Dennis B

Preliminary 2012 data from the national General Social Survey have just been released, and if the numbers hold, they show an “all time high” for public support for space exploration program spending, with positive responses (“too little spending”) higher than at any point in the survey’s 40-year history, and negative responses (“too much spending”) lower than the post-Challenger and early Apollo periods.

I’ve previously written about the GSS data (see here and here for more details on the survey, and here for some trends beyond space exploration).   The 2012 updates below use preliminary data rather than the final release, so handle with care:

Initial signs of an upward trend showed up in 2010′s “too little” responses, but were accompanied by a rise in “too much” responses as well, suggesting a polarization.   The “too much” trend has reversed itself in 2012, leading to an overall higher favorability rating stronger than at any point in the combined GSS (1973-2012) and Launius (1965-1971) history.

A slightly different rending shows the composite scores (note, “don’t know” responses are not included here):

Finally, the GSS question variant which excludes the word “program” shows the same trends (“program” tends to yield slightly less favorable responses, presumably due to a baseline anti-government-spending sentiment):

Stay tuned for updates using the full weighted and quality controlled data when they are released…

Universal Principles of Design: 100 Ways to Enhance Usability, Influence Perception, Increase Appeal, Make Better Design Decisions, and Teach Through Design by William Lidwell
My rating: 5 of 5 stars

I can’t say enough good things about this book. Beautiful and inspirational, it has the reassuring feature that a book on foundational principles of design should be exceptionally well designed itself. Each principle gets a two page spread with just the right depth of summary, narrative and citations on one side, and exceptionally well chosen examples on the other.

The principles bridge graphic design, industrial design and engineering and can greatly aid building a common language of design and better understanding and collaboration in organizations which contain these often not-well-married communities. While the principles themselves might range from pithy to thought provoking, the overall ensemble is very rewarding.

Hands down, 5 stars.

View all my reviews

The Signal and the Noise: Why So Many Predictions Fail – But Some Don’t by Nate Silver
My rating: 4 of 5 stars

As with some others on Goodreads, I found this book a little hard to rate, thinking it a “3.5″ and opting for a 4 star rating from an “E for Effort” standpoint. Part of this is high expectations on my part based on affinity for Silver’s FiveThirtyEight election prediction work.

The book is well researched and covers a nicely diverse array of example topics, including but not limited to economics, betting, sports, weather, climate, earthquakes and terrorism. The diversity keeps the interest going. A challenge here is that few of the examples were unfamiliar to me; ironically as the book is ultimately about Bayesian inference, there may be a little bit of a Bayesian thing going on relative to those most likely to buy/read and those most likely to have prior exposure and be left wanting more. The same thinking might suggest that the book is targeted more towards readers attracted by Silver’s political forecasts than those with a wonkish or professional interest in prediction itself.

For the latter, Silver redeems by offering something hard to find in similar popular literature, a high level synthesis across both realms and disciplines in prediction. A contrast with Kahneman’s Thinking, Fast and Slow and Surowiecki’s The Wisdom of Crowds (both of which Silver draws from) helps illustrate: While these two books are by no means peers (Kahneman’s represents a lifetime of scholarship, Surowiecki’s is more management faddish), as books, both suffer a bit from “the curse of knowledge” – the authors’ over familiarity with the often contradictory details leaves the reader rudderless on how to apply the findings in practice.

Silver, instead, takes a first step towards synthesis. This is welcome, although occasionally questions do arise about the formal correctness of mixing and matching themes and findings from very the different predictive methods (regression, classification, physics based modeling, simulation, etc) covered in the book. Absent a unifying framework to relate these methods (Silver is clearly an applied forecaster rather than a theoretician) the reader must rely on his claims to authority by experience (as well as the depth of research indicated by heavy citation) in trusting the synthesis and recommendations.

Overall, Silver ends up on the positive side of the trust ledger sheet, and even for readers already familiar with the topical examples, he provides enough additional color, as well as thought provoking commentary, to make it all worthwhile.

View all my reviews

Another Science Fiction: Advertising the Space Race 1957-1962 by Megan Prelinger
My rating: 5 of 5 stars

Wow! This book is not only beautiful but really interesting. Far more than just a nostalgia coffee-table piece, the text is both interesting and thought-provoking.

I found the best chapters / collections to be those on the human form, and on modern art influences. In these not only were the graphics themselves stunning, but the narrative added significant depth. I was a smidge less interested in the chapters that spent more time writing about technical details – the content seemed too deep for a non-aerospace reader, but “already known” for insiders.

A surprising side to this book was how many of the advertisements are focused on recruiting, reflecting the early space age ramp up. To this end the ad copy itself is often very interesting as well.

I gather the author is hard at work on another book. Can’t wait!

View all my reviews

The Idea Factory: Bell Labs and the Great Age of American Innovation by Jon Gertner
My rating: 3 of 5 stars

Jon Gertner’s “The Idea Factory” tells an important story about the history of many of the communications and information technology underpinnings of our current era. More importantly, it explores (indirectly and eventually) a major question of what is needed to make large basic and applied research labs successful. I’m glad I read this book, but can’t say I necessarily enjoyed reading it. As such I’m struggling with whether to rate 3 or 4 stars … if Goodreads allowed 3.5, that’d be it.

Growing up very close to Bell Labs’ Holmdel NJ facility, I was attracted to this book because of the place the Labs occupied in our local culture. If you were bright, technically oriented, and wanted a well-paying job, Bell Labs was the place to strive for. The invention/discovery of radio astronomy at Crawford Hill added to the mystique.

The book’s primary drawback (for me) is what feels like a 70/30 split on biography versus technology – I think it would have been a more interesting and engaging narrative had the biographies of a few famous players in the Labs’ long history been treated as embellishment, rather than primary narrative. (I say this regretfully as many of these individuals and their world-changing contributions deserve recognition, but as a book review, it makes for a challenging narrative structure). The reality is the connection between the sometimes (but not always) colorful individual personalities and the ideas and technologies they generated is often too thin; the researchers emerge as characters in a narrative without a strong or clear connection to the plot devices, so to speak. Claude Shannon, whose prescience in 1950′s communications/information theory was truly remarkable, is perhaps an exception.

In short, the secrets to success of the Bell Labs miracle, which generated an amazing number of globally transformational technologies, seem to fall into:

1 – access to the huge reservoirs of working capital that the AT&T monopoly and “cash cows” (local telephone companies) to fund large scale R&D

2 – protection from competition, insofar as it created an environment where fundamental / basic / applied research could be pursued as an end of itself (shielding researchers from having to seek funding), working towards 30-year (not product cycle driven) time horizons, and loosely organized and driven by :

3 – the “grand challenge” of emplacing and continually improving the nation’s telephone network, at its creation, the most complex “machine” (or at least system) ever considered

4 – the co-location of the basic/applied research function with a very large engineering and systems development capability, which could turn fundamental discoveries into applications

The pregnant question – treated only glancingly in the final 2-3 chapters – is whether such a construct is possible again in today’s world. The industrial lab model (at least at this scale) seems to have been supplanted by the venture/entrepreneurial distributed model, but as Gertner points out, this has a defect in that it sacrifices item #2 above, driving innovation to be more incremental and product driven, than transformational. It is unfortunate that this theme was not more deeply developed in the book, as well as a deeper treatment of the evolution of Bell Labs after the AT&T breakup of the 1980s; this is rushed through in single chapter, and yet it is the crux of one of the more thought provoking and relevant lines of thinking. Also only glancingly treated are whether information technology (for which the only giants capable of investing such R&D today – Apple, Facebook, Google, etc) is even the right domain, or whether the next grand challenges will arise in biology or energy.

As a long-time NASA employee, I can’t help but try and draw contrasts between the Bell Labs 4-point model summarized above, and how the Agency currently operates (recognizing that this is peripheral to the book review – but it does illustrate how the book’s concepts have relevance):

1 – At its face, the “capital sufficiency” test is met, with an $18B/year annual budget, NASA is exceedingly fortunate relative to government R&D agency peers. However, the budget is profoundly oversubscribed, and broken down (with significant administration and legislative branch ‘assistance’) into a number of stovepipes which do not intercommunicate or leverage resources well. As a result, basic and applied research are confoundingly “resource starved” at the lowest levels.

2 – Competition is a mixed bag. As a government agency, NASA must embrace competition among researchers as part of its fiduciary responsibilities to taxpayers. That said, competition between NASA ‘business units’, as well as the machinery of a grants process, tends to disfavor basic and applied research for its own sake, working to objectives with much longer time horizons than the needs of individual programs and program managers. The “luxury factor” AT&T was able to provide as a corporate culture – and management – decision isn’t within the purview of program managers to grant.

3 – While NASA faces many challenges in doing some of the most complex engineering and discovery possible, it (1) lacks the “grand challenge” incentive that drove emplacement of the national phone system, and (2) has for too many decades been hostage to warring philosophies – and politics – over purpose. It may be that until a truly globally critical function becomes apparent – energy or rare earth metal supply, planetary protection, space security – a driving purpose as vast, technologically provocative, economically relevant and complex as the phone system creation, may be elusive.

4 – This piece of the Bell Labs equation is present in NASA, co-location (at least at an Agency level) of a basic/applied research base with a vast capability in engineering development. Unfortunately, the capabilities are too stovepiped, with the research capabilities and centers too poorly integrated with the development capabilities and centers (as well as programmatic impedance towards making the transition). Of the four factors, this is the one that would be most readily addressed by changes in management approach (although by no means easy).

Coming back to the book review – I would say The Idea Factory is a “should read” for those interested in the topic of innovation or R&D rather than a “must read”. It is rewarding, but only with the investment of a good bit of energy, patience and passion for the subject.

View all my reviews

The Theory That Would Not Die: How Bayes’ Rule Cracked the Enigma Code, Hunted Down Russian Submarines, and Emerged Triumphant from Two Centuries of Controversy by Sharon Bertsch McGrayne
My rating: 4 of 5 stars

It probably takes a special sort of person to dive into an entire book about one statistical theory, but for those so-motivated, this one pays off.

The pro’s: The author has done a phenomenal job at capturing and richly detailing the very “large” personalities that have championed (or condemned) the use of Bayes’ Rule through the centuries, amidst a little-known and long-simmering war that has persisted between statistical Bayesians and frequentists since the concept was first brought forward. This is even more impressive as she is a journalist, rather than a statistician. McGrayne immerses the reader in what can only be called “lush” detail of the history, from personalities to global events.

The con’s: This a very dense text. Not dry in an academic sense, but a lot of material to consume. At times I had to summon extra reserves of motivation to proceed to the next chapter. The topic is also a difficult one to communicate solely through narrative – more than once I found myself wishing for just a little bit of math-by-the-way-of-example to help grasp the concepts. (With such, this could actually serve well as an educational vehicle). While already familiar with Bayes, the application in some of the historical examples was, for me, elusive.

Computing power has today made the Bayesian/frequentist conflict somewhat moot, and I found myself wishing for a little more exposition of Bayesian applications in the modern era. (To me, this is where the real excitement lies, if “excitement” is the correct term!)

Overall – if statistics, scientific inference, decision theory or machine learning excite you, this is probably a book to have under your belt. Reading the history of Bayesian vs frequentist wars triggered some good musing and reflection on the critical question of “how to make inferences when too little, rather than too much, data are at hand”.

View all my reviews

As a latest foray into lifelogging I’ve been playing with a tool called “Moodscope“, a lightweight daily quiz which tracks excursions into positive and negative moods.  It appears to be rooted in the established psychometric metric of the Positive and Negative Affect Schedule (PANAS).  In short, you take a short quiz, and get back a 0-100 score, with 50 being neutral, >50 being “positive affect” (er, “good mood”), <50 being “negative affect” (“bad mood”).   Faced with such a temptation, I couldn’t resist seeing if this was all hooey or if there would be signal in the noise…

To cut to the chase … much to my surprise, I could actually track long term trends.  Shown at left are both my daily and weekly-averaged data for the last three months.  Even without the weekly averaging the ‘envelope’ of the raw data suggested trends.   (I also have a pretty good intuitive understanding of the related causes, but there’s only so much I’m going to dump onto a public blog!)

Since much of my mood is driven by what’s going on at work (both negative, on the frustration side, and positive, on the “things that feed me” side), I next wondered if there were any weekly trends buried in the data.   Switch the smoothing filter from 7-day to 3-day and … voila! :

There’s the expected weekly cycle, hiding out in the raw data.   While I hypothesized there might be such a signal, I never anticipated it would be that strong.   But in hindsight it makes sense, my work-weeks, while not routine, do tend to share some general characteristics week-to-week.  Switching to a distribution-based view of the raw scores, the story becomes, “Tell Me Why I Don’t Like Monday’s” (that’s an easy one actually; it’s all staff meetings, ‘sync tags’, etc, with very few opportunities to interact with people in motivating or inspiring ways – all activity, no energy).

(For stats junkies, the indented boxplots are a cool graph the Mac stats program Aabel generates … fatter boxes at a given score mean more occurrences; diamonds show the mean value, solid horizontal bars show the median value).

Anyway, back to the data.   After lousy-Mondays, I typically get to work a lot more with people on, well, actually creating stuff, rather than managing it.   Whether or not I’m doing it myself or just coaching, this is what fires me up.   Of course, then the week trudges along, energy gets drained, stress goes up, backlogs grow, opportunities for conflict increase, etc, leading to not-so-pleasant Fridays.   By the weekend, I’m mostly in “recharge” mode (“meh”), hovering more around neutral values.  Unfortunately the PANAS positive affect questions all seem to be skewed towards very “active” positive values (“determined”, “inspired”, “interested”, etc); there’s no way to get positive affect credit for things like “cozy”, “content”, “relaxed”, etc).

As a final exercise I took the weekly data and simplified it down to a simpler user’s guide; for any given day of the week, how likely is it that I’m going to have a good day (mood score > 60), a bad day (mood score < 40), or a “meh” day (40-60).   Arbitrary thresholds, I just picked ones that divvied up the data well.   And yes, these are now posted on my office door (how’s that for biofeedback?!)

Pausing for a “work meets play” moment, thanks to stumbling across the cool interactive timeline tools at tiki-toki.com.   To try out the new tool, I loaded up a database of NASA mission proposal competitions over the last 20 years, mostly in science and technology.   The dry run is available here.  (Or view a screencap here).

Time for one last slice of the NASA pie, this time broken out state by state (following up on previous slices by Company, and by Field Center).   Again, all data are from NASA’s annual procurement reports, are inflation-adjusted to FY10 dollars, and are 3-year smoothed.  The embedded time-series begin in 1999 (reflecting a 1997-1999 average) and end in 2010 (thus, the data still only reflect the old Constellation program, and not the Obama administration’s requested new direction).

On the “large and growing” side, there are few surprises; Colorado stands as the biggest “winner”, due to Lockheed Martin’s capture of the Orion / MPCV crew vehicle.  The trends of the solid rocket industry are largely captured in the Utah data; the recent rise reflects development of the 5-segment reusable boosters slated for the new Ares rockets, prior to their cancellation by the administration.  Florida’s rise over time is likely related to procurement of launch services from the EELV fleet of Delta and Atlas rockets, mostly for scientific missions.

On the shrinking side, Louisiana’s decline is not surprising, as production of the Space Shuttle’s large external fuel tank slowed, then ceased.  The nearly $1B drop in Texas spending may be related to completion of the development of International Space Station modules and systems, in the earlier part of the last decade.

As notable as the ‘winners’ and ‘losers’ is the more than 40 dB (4 orders of magnitude) difference between states receiving the least, and the most, NASA direct spending.  (Emphasis on direct, since these data only reflect NASA “prime” and first order service contract procurements.   A web of subtier supplier industries provide services to these contractors, so the actual national distribution of NASA dollars would be slightly more uniform).

Here is the PDF version of the above graphic:  PieSlice_States

Last week I sliced the NASA procurement history up based on its deployment through its ten field Centers.  Today, a slightly different take – tracking the end recipients.    Well, to a point … NASA’s official procurement data only tracks the “prime contractor” recipients of awards (whether technical primes or support service contractors).   In the actual aerospace industry, of course, the money then gets distributed through layers of subtier suppliers (subcontractors).   That said – even tracking the primes over time is interesting.

The visualization below shows the trends for selected major contractors from 1997-2010.   To construct the graphic, the following simplification were made:

• All data are converted to FY10 dollars; all inset numbers are in $M per year
• Previous 3-year smoothing is applied (e.g., the year 2000 data point is the average of the 1998, 1999 and 2000 point data)
• I have tracked industry mergers to the best of my ability and report only the “latest standing” prime contractor … an incomplete list follows the post
• I’ve excluded a number of primarily construction contractors, as well as many smaller contractors that were active in the late 90′s / early 00′s, but are not active today.

Today’s post is just a data rack-and-stack-and-dump; this time of fifty years of NASA annual procurement report data.   (Add NASA to the long list of Federal agencies which report annual data, but do not trend it).  As with many such data sets, a little graphic visualization often helps understand long term historical trends.

First, the procurement data broken out by field Center.  In all plots below, historical real year dollars have been inflation-adjusted to FY10 dollars, and each plotted year corresponds to the average of the previous five years’ procurement data (the data are smoothed).

The trends are a little more coherent if grouped by major field Center ‘type’, i.e. Human Space Flight (Marshall, Stennis, Kennedy, Johnson), Aeronautics/Research (Glenn, Ames, Langley, Dryden) or Science (Goddard, Jet Propulsion Lab):

In this view, the slow but steady increase in “procurement share” of the science centers since the early 1980′s stands out.   Human spaceflight centers have accounted for a fairly stable 50-60% of Agency procurements since the early 1970′s.

Rendered as curves the trends are also apparent, modulated by the long term trends in the Agency’s budget:

Finally – and less expectedly – the share of NASA’s spending that occurs as procurements has interesting trends.   Almost all of NASA’s budget occurs as procurements – i.e., its services are bought from industry, in the form of prime or support contracts.   This has been true since its inception.   The interesting trend occurs from 1990-present though, as the procurement percentage has steadily dropped from 90% to about 83%:

Since there was no 20-year acquisition policy which inclined towards insourcing over this period, the most likely explanation is that this reflects the steady increase in Agency fixed costs, relative to the total budget.   These fixed costs might include the costs of aging infrastructure, as well as utilities and civil servant salary expenses which may have increased faster than both inflation and the Agency’s top line budget.

NASA enjoys the stablest budget in the Federal government.  Why then its continual state of disarray?

To explore this question, I’ll start with a refresh and update of “Ask and You Shan’t Receive”, where I riffed a NYT “porcupine graphic” to illustrate the differences between Presidential budget requests, and actual Congressional appropriations, for the NASA budget.   I’ve added in some earlier budget years’ worth of data, and turned the crank for DoD as well.

First, an eye-opener from the defense data.   I’ve previously illustrated the magnitude of long term defense cyclicals, which are fairly large in actual, appropriated amounts.   Even more impressive (disturbing?) is the disconnect between administration / defense department long range (5-year) planning, and the actual spending amounts:

After 2003, part of the disconnect is attributable towards appropriating the wars in Iraq and Afghanistan outside the baseline DoD budget request, year to year.   However, even prior to that, it was not uncommon for the out-years of defense plans to be hundreds of billions of dollars out of bed with the eventual, actual appropriations.  That’s a lot of plans that never made it to fruition (yet somehow, the nation’s defense capability remained and remains intact and operational).   To call the defense planning and budgeting environment “volatile” would be an understatement of the highest order.

An update to the equivalent NASA data shows a much more “tempered” situation:

With the exception of the 1990′s (begun with ambitious requests for Space Station Freedom and the Space Exploration Initiative; ended with aggressive plans for downsizing of NASA), there have been only comparatively minor differences between administration out-year planning and congressional appropriation.   This has certainly been the case through the 2000′s.

To drive home the point further, the same data are integrated over 3-year windows and overlaid for both DoD and NASA:

Further, NASA’s overall budget fluctuates significantly less than most other Federal agencies, year to year.   Comparatively speaking, NASA’s funding is the stablest in the Federal government:

We thus have a situation where:

• NASA’s actual year-to-year budget is the stablest in the Federal government
• Administrations and congresses tend to agree on the right “size” of the Agency in any given year, far more so than DoD (and by implication of the variability above, likely more than other agencies as well)
• NASA’s programmatic portfolio is nonetheless continually in disarray, both in planning and execution

The last bullet is particularly confounding.

Over the decades various blue ribbon panels have repeatedly exhorted that “NASA is asked to do too much with too little”, however, this rationale is likely true of every Federal agency.   The “finding” is both unsurprising (coming as it does from inside-the-bubble panels) and not particularly helpful; providing advocacy, not advice.

Others observe that since Challenger, NASA has found itself repeatedly trapped between severely divergent, and seemingly irreconcilable, ideological and political belief systems about the next steps in the development and exploration of space.  Aligned with powerful legislative and industrial stakeholder interests, the shuttlecock of civil space bounces back and forth.   Unfortunately, this again is true of most sectors of Federal government.   It suffices as a description for the situation, but fails as an excuse.

Some of this disarray must come home to the Agency itself to roost.  As an agency within the executive branch, it is NASA’s job to balance competing stakeholder interests, fluctuating budgets, midstream replanning, and future uncertainties.  This job is neither unique, nor more difficult than in any other branch of the Federal government that engages in long term development (rather than services or operations) programs.   Put simply, it goes with the turf.

What NASA needs first and foremost, I will argue, is neither new vision, nor the “right” 20-year plan, nor revolutionary innovation, nor bottoms-up reinvention.

Rather, it needs to continue making progress on a long term path towards disciplined management.

As a first step, the Agency desperately needs to re-establish the capability to formulate and implement coherent, integrated, and self-consistent budgets (briefly grasped during the early years of PPBE implementation, but lost more recently for a variety of reasons).   The challenge here will be to restore discipline, while not succumbing to the lure of central planning, since…

… as a second step, the Agency must develop the ability to adjust these budgets rapidly, and flexibly, in response to the winds of change on both sides of the Mall.  Currently, minor (a few percentage points) adjustments to NASA’s budget lead to crippling paralysis in the system.  This is an inevitable outcome of attempting to manage detailed implementation outcomes at too high a level (a trend which continues unabated), and pursuing too many conflicting agendas simultaneously.  Failures to delegate appropriate decisions to appropriate levels, or to manage by budget, are exacerbated by a field center culture which has evolved to treat the institution itself as an entitlement, rather than a tool.

At the end of the day, the budget data speak loud and clear:  Relative to other Agencies, NASA is blown by a light breeze in the eye of the Federal budget storm, not Cat-5 winds at landfall.   As an Agency it has enjoyed – and continues to enjoy – a level of budgetary stability almost unprecedented in the Federal government.  The “complexity” of its situation is ultimately tempered by that simple reality; rationales for failing to quickly adapt to change are few.  The budgetary “safe zone” has been ours to either use, or to abuse and lose.  On our current vector, the risk of the latter seems extremely high.

Selected References

OMB Historical Budget Tables (and Deflators)

Department of Defense Green Books

Older DoD Future Years Defense Plans (see Table 14)

Older SecDef Annual Reports

NASA Budget Estimates

More NASA Budget Data

Here are some snapshots of local color during the fantastic EyeO Festival held in Minneapolis earlier this week… I cajoled (little arm-twisting required) a couple of friends into joining in on a field trip / vacation to this event.   Call it a “mental spa weekend”.   Having never been to Minneapolis before I have to say I was very impressed … the city has incorporated a great artsy vibe, and has done a magnificent restoration of the mill waterfront district.   Even the “Hour of Flour” at the Washburn / General Mills museum was more interesting than I thought.

The photographs below are mostly “obvious” targets but nature was very cooperative.   For the faded Pepsi sign shot, I was lucky enough to find reflected sunset/golden hour light, off of the glass facade of a large building across the street (this wall is eastward facing).  The Murray’s photo I snuck in just because it’s so wonderfully font-o-rrific.  Doors-to-nowhere are obligatory catches, and transformers are just plain cool.   The evening shot of the Mill museum and commercial loft spaces isn’t very well done technically, but captures how beautifully the city has merged old and new in the Mill district renovation.    Enjoy!

Last week I shared some findings on how the new GoogleCorrelate tool uncovered the very sharp seasonality in searches for student internships (happily and coincidentally helping me out at work).   Search interest follows an extremely concentrated annual cycle, beginning during (or just after) winter break.   This of course begs the question of what topics are of most “peaked” interest for the remainder of the year…

With a few (52) manual iterations of adjusting the offsets in the “Winter Wave” pre-programmed GoogleCorrelate search, I’ve arrived at the following visualization.  The outer ring denotes the “most peaked” term for each week of the year, the middle ring denotes the “second most peaked” term in the lists, and the inner ring denotes the “third most peaked” term.   Poking around on the actual GoogleCorrelate site, there are some really interesting nuggets buried deeper in the rankings, but with a flat graphic I’m limited by how much data I can cram in!

The real “Winter Wave” is almost a little comical.   After shopping for designer boots, the jet-setting crowd sets off transworld snowboarding, for which they need Nordica and Volkl ski equipment.   After the holidays, the fun is a little closer to home (hockey arenas), after which all this extreme exposure results in much more mundane concerns (oranges, presumably to help cure bronchiolitis).    Well, OK, that may be more of a coherent narrative than is justified by the data, but it’s nonetheless fun.

The February-April rat race of summer internships, summer programs, and finally summer camps makes me a little glad to be neither a kid, nor a parent, in the 21st century.

The summertime drill from May-June appears to be:

• Dress to kill (leather sandals, linen clothing, born sandals)
• Fix-it-up (awning, decking, cedar decking, deck railing)
• Keep-it-at-bay (Brush Hog, trumpet vine, johnson grass)

If you’re selling pre-packaged christmas musicals for kids, get them on the shelves by early September.   October and November are completely devoted to killing things with 4, rather than 6-8, legs (treestands, duck blinds, deer stands, upland hunting, waterfowl).

After which we go shopping for designer boots and begin it all over again…

A while back I published a hard look at the optimistic launch demand estimates generated by FAA’s COMSTAC annual forecast.  This year’s forecast has just been released.    In addition to the forecast, the report has for many years included a valuable and underutilized source of information, which provides a more believable and fundamental basis for optimism in the commercial launch sector.  

The Data

The data are an annual survey of satellite service providers on market, regulatory and industry factors which either promote, or inhibit, the service sector’s demand for new satellite launches.   While the data are presented each year, I have never seen them actually trended anywhere.   Doing so provides some good evidence that the market may indeed by looking up for commercial launch providers, independent of the self-referential and squirrelly launch demand forecasts.

While the sample size is small (this year, 14 companies responded), the coherence of the results trended over time indicates that there is some value to the data.   Each of these companies were asked whether various factors had “significant negative”, “some negative”, “no effect”, “some positive” or “significant positive” impact on their plans to purchase and launch satellites.   The factors included the readiness of the launch and satellite service base (launch vehicle and satellite availability and reliability), the financial and regulatory environment (economic conditions, availability of financing and insurance, and ease of securing operating and export licenses), and finally structural changes in the overall market and competitive landscape (demand for services, competition from other service providers, competition from ground-based providers, and increasing lifetime of satellites).   Note that these are my three “meta-categories”, rather than COMSTAC’s.    Trended over time (and aggregated), the results are very informative:

The schema here is that dark red, orange, yellow, light green, dark green correspond to the percent of survey responses in the corresponding “significant negative impact” to “significant positive impact” categories above.  Even with a small survey sample (with variable participants from year to year), time trends are coherent and clear.   The self-reported perception of these large providers is that overall conditions have indeed improved significantly since 2003.

The Good and the Bad

What is most interesting is which factors have improved, and which are still the sticking points.   The readiness and availability of both launch and satellite services has evolved from being a primarily “no effect” issue in 2003 to a net positive factor in 2011.  This is strongly driven by improvement on the satellite side, although slow improvement on the launch side contributes as well.  In 2011 no respondent rated launch vehicle availability or reliability as a “significant negative” impact.

Rather, the remaining key issues all line up on the “business side” of the equation.  Financial and regulatory issues were ranked a significant burden in the early 2000′s.  Since then it appears that a supportive financing, insuring and licensing infrastructure is beginning to emerge.   The limiters in this category continue to be export licensing, as well as global economic conditions (and the closely linked category of financing availability).   Should the economic environment improve, and ITAR issues be addressed, this category could evolve to be at least “neutral”.

Inside the Fight Cage

Within the category I am calling “market changes”, the most notable trend is a steady increase in overall demand for satellite services, ranked strongly as a positive or significant positive factor.   The other factors in this category are structural within the industry and market, relating to competition, consolidation and innovation.   The overall picture is one in which demand for services is high (and increasing) but the viability of any given provider’s position to provide those services within the market competitively (and from space) might fluctuate.   When combined with innovation (increasing satellite lifetime, and hence downward pressure on demand for hardware and launch), this category overall remains a mixed bag of impact for launch providers.

The data seem to provide a clear and intuitively acceptable story of evolution within the sector over the last ten years.  Unfortunately, we lack critical validation data, since the survey in its current form was initiated after the collapse of the late-1990′s commercial services-and-launch bubble.   Reference data from a comparable questionnaire during that period would have gone a long way in helping to calibrate out unfounded optimism.   The best we can point to is the fact that many of the subfactors point towards reasonably objective “externalities”, such as licensing, financing and insurance availability.   Further confidence might be found in the sheer size of the market in play, on the services side – at least two orders of magnitude larger than that of the launch industry, as per the latest annual Space Report.   The idea is to tune in to the folks with the biggest stakes in the game (the space products, services and infrastructure survey respondents):

Whatever the details of the breakout (including fine details around commercial vs non-commercial launch) it is clear that the center of gravity in the space economy is on the demand (services) side, and it is that side that should be focused upon when assessing the climate for future launch activity (instead of the supply side, i.e. launch providers themselves).

Put simplistically: if the fluctuations of an existing $189B commercial space products and services market, and an existing $86B governmental services market, have been unable to cause more than minor fluctuations in an essentially flat rate of actual launches per year, it would be imprudent to assume that fundamentally new markets predicted by emergent launch providers are likely to change the overall demand picture dramatically (or at least quickly).

Kool-Aid vs Protein Drinks

This is not to say the outlook isn’t optimistic, but the optimism should stem from much more mundane (and sustainable, and believable) factors than anticipation of self-bootstrapping “airmail miracles” of new market generation.  (There, I’ve lost 50 “Fast Company” brownie points with one simple sentence.   So be it).   The gains reported in the surveys are much, much more prosaic.   Satellites (rather than launch vehicles) have crossed a reliability/availability threshold and are now on average seen as positive factors.  The operating environment (both financial and regulatory) is increasingly supportive (or at least, much less inhibitive) of a space sector economy.   Finally, demand for services as a positive factor is improving much more rapidly than negative structural changes within the competitive environment seem to be offsetting it.   All of these point towards strong fundamental (rather than speculative) causes for mid-to-long term improvement in demand for launch services.

Postscript

A final note on forecasting: eyeballing the composite survey trends above, versus the actual launch rate data for the last decade, it looks like a regression-type actual forecast model driven by the survey, or the survey plus recent actuals, might be possible.   This would be limited by the number of available data points (currently 8, from 2003-2010) but wouldn’t have to incorporate much skill to outperform the actual COMSTAC forecast, which I think (still to validate) has less skill than a “persistence forecasts”.   Stay tuned for a future post on this topic…

Want to attract the attention of the best summer interns?   Target your communications campaign to be ready on the first of the year.   It shouldn’t be a surprise that interest in internships follows an annual cycle with a peak in the late winter and spring … but what it is surprising is how sharply peaked this interest is.

Two separate threads led me to the results below: (1) Experimenting with Google’s new experimental GoogleCorrelate tool, and (2) helping Marshall Space Flight Center‘s Summer and Minority Internship Programs improve their external communications.  GoogleCorrelate adds a new twist to the analytics mining provided by Google Insights; it allows “reverse engineering” to find which search terms most closely match pre-defined geographic patterns (e.g, state-by-state demographics), or pre-defined temporal patterns (e.g., an annual yearly cycle).

The time search, driven by a simple annual sine wave in Google’s Winter Wave example, identifies which terms are most likely to be searched with a pronounced annual peak.   By adding an offset value, the peak can be shifted in time.   Playing around with the tool, I was surprised to find how strongly summer internships, summer programs, etc, dominated the searches for all peaks centered in the January-to-March time frame:

Switching over to GoogleInsights allowed plotting and extraction of the actual annual data in CSV format (the first step in generating a composite).   To increase the data content, I extracted both “summer internship” and “summer internships” (plural):

After a little bit of smoothing, we get the average annual cycle of summer internship searching:

This is probably nothing that education specialists don’t already know, but may not known by our supporting communications staff.   The message I take away is clear: kids (or parents) start searching during and immediately after Christmas break, with peak interest the first two weeks in January.

From a communications perspective, any data which point to a more-than-doubling of target audience interest over a narrow window are of great interest.   The associated promotional campaign recommendations would include making sure updated supporting content or collateral is firmly in place before the holidays.   January would then  make an excellent month to reinforce interest and enthusiasm by highlighting the internship program in executive speeches, mobilization of intern program alumni or targeted college almuni already in the employee ranks, etc.  Even small slips in the communications campaign schedule could lead to significant gaps in reaching the “first wave” of internship seekers, who may also be the most motivated.

Earlier this month I posted an analysis of which countries most need Kiva.org’s microfinance services (based on their poverty rates) and which countries are getting them (based on Kiva’s actual lending data).   This analysis provides an opportunity to update my ranking of Kiva in-country microfinance banking institutions from last fall, to include “relative need”.November’s ranking derived from three factors: an MFI’s interest charging practices, its cumulative experience, and its risk level.   The ranking simply composited these factors.  A downside of this approach is that some of the neediest countries – e.g., many if Africa – by their nature may cause MFI’s to score poorly in the rankings.   (For example, poor infrastructure such as roads, etc., will drive up the interest rates MFI’s must charge, to cover the transactional costs of the loans.)

Including the “need factor” based on poverty rates and Kiva penetration helps offset this bias, and helps identify those MFI’s that may be “worth taking a chance on” in high-need, but higher-risk, countries.   Below is a revised regional ranking, in which I have also filtered out about a third of the MFIs with extremely low scores in any of the four metrics:

To simplify things, I’ve also recreated a “top 20 table” from the overall rankings.   In this table, any MFI with a score in the lower third of any of the four sub-factors has been filtered out.   In short, this table provides a way to pick Kiva loans such that your loan dollars go to the neediest regions of the world, while balancing your financial risk and rewarding institutions that keep their interest rates within (or better than) the norm for microloans.

It’s been a while since Part 1 of this photo journal, but I’m finally getting around to rounding it out.  Here are ten more studies of photos that almost – but didn’t quite – make the mark … but are still quite a sight!

The shots below are all from 2010, when I spent a few days in July traveling around the French province of Aveyron, thanks to a wonderful tip from a friend.

Wildflowers at Najac

Through the ruins at Najac

La Bastide des Fonts, near Cornus

Atop the chateau at Belcastel

Grasses in southeast Aveyron

Treed texture in Brousse-le-Chateau

 Along Le Cernon, near La Bastide-Pradines

Aveyron Blue

Open for business?

Obligatory photo of the Milau Viaduct

Bain & Company has released the 2011 data from their biannual Strategic Management Tools survey, so I’ve taken the opportunity to update my “meta-analysis” of these tools,  a mashup of the full 20 years of Bain data with Google Books and Google Insights trending.   The result, the Strategic Management Tools Fad-O-Meter 2011, version 1.1.   Without further ado:

The Freshness score is computed as in the original post; Freshness is a composite of the trending upwards (or downwards) in Bain Survey Utilization scores (1993-present), Google Books (1990-present) and Google Search Insights (2004-present).   The scores are normalized to percentile rank, 0-100.   Again, the old touchstone TQM shows that tools don’t have to be fresh to be effective, especially when honed towards their target niche applications.

Satisfaction is also computed as in the original post, it is a composite of the Bain average satisfaction score for the past 10 years, as well as the satisfaction trending direction (thus, “new entrants” which have low absolute satisfaction but are trending upwards, have a chance to stand out, while “new entrants” which arrive with a bang but with experience lose their luster, also get corrected).  Because this score composites both the mean and the trend, it is important to cross-compare against the absolute satisfaction score at the bottom of this post (as an example, even though strategic planning leads all tools in absolute satisfaction, it is ranked as “middle of the pack” in the Fad-O-Meter since its satisfaction trend is nonetheless slightly downward … we seem to have run out of ways to squeeze even further performance from this old touchstone tool, and may need new tools for today’s business environment).

Tools which focus on knowing “yourself”, your competitive environment and your customers occupy strong positions in both buzz and satisfaction (core competencies, vision statements, customer segmentation, benchmarking, scenario planning).

Open or collaborative innovation tops the freshness (hype) chart right now, but has yet to demonstrate itself in Bain satisfaction scores; the case is similar for supply chain management.   After many years, balanced scorecards maintain their buzz but are similarly lagging in generating satisfaction among executives.    Overall, tools which focus on structural or process change (or management) rank either in the middle of the pack, or as laggers, in satisfaction (TQM being a notable exception).

In the Bain 2×2 format (power-boosted to 3×3), here are 10-year composites for both usage and satisfaction (averaging helps bring out continuity across the biannual Bain surveys).   This helps separate (as Bain analysts note) “power tools” (high usage, high satisfaction), “niche tools” (low usage, high satisfaction), “blunt instruments” (high usage, low satisfaction) and “watch items” (low usage, low satisfaction).

If you’re just interested in executive satisfaction, here are the ranked 10-year average satisfaction scores (i.e., reading from right to left on the chart above):

Mapping meets microlending, two of my favorite topics!   Having previously ranked the quality of Kiva microloan partners (and now, for several months, having used the rankings to steer my own loans), I thought I would do some quick and dirty visualization of the results.  As a bonus I’ve churned out some visualizations of worldwide Kiva lending, poverty rates, and the relative penetration of Kiva into the neediest countries in the world.

My default quick-and-dirty visualization workhorse is IBM ManyEyes – lacking in design elegance and configurability, but free, interactive and very versatile.  Whether in color density or bubble mode (click the link to try that out), I’m a little underwhelmed by the results.   I’ve only got 35 loans to work with, so the starting data are thin, but nonetheless, the graphic looks a little too spartan for my tastes.

I’ve also tried a Kiva dashboard page using Nick Felton’s Daytum service (at right); slick but more of a curiosity than anything else.   (Nick and Ryan are now on their way over to work for Facebook, but I’m still holding out hope of WordPress-embeddable Daytum panes, and some point!)

Visualization-wise, my best results came by simply adding locations to GoogleMaps, along with judicious selection of a good contrast background (the terrain map).  This does (hopefully) a better job at pulling out my preference towards funding women borrowers, showing my interest in lending to Mongolia (spurred by both a visit there, and the admirable social lending practices of XacBank, including both microsavings reinvestment, and green loans) as well as several countries along the former Silk Roads (derivative Central Asian interest from the Mongolia thing, plus some empathy from the locals for surviving millenia of cross-cultural political intrigue).

Of course, my own loans are of interest to, well, no one but me.   More interesting is the distribution of all Kiva microloans to date.  Most interesting is this distribution compared with the worldwide occurrence of poverty.   For the latter, I’ve pulled data from the World Bank, specifically, country-by-country percentages of population living on $2 (USD) per day, or less.   Combining that with raw population data converts percentages to poor population estimates – the upper panel in the graphic below:

Clicking the graphic will take you to the live visualization at ManyEyes.org (where you can also visualize many other parameters).   In the lower panel above, color intensity shows where Kiva lending has made the “biggest dent” in various countries’ poorest populations (Kiva dollars loaned per “poor” individual).  The imbalance between microloan need (upper panel) and Kiva supply (lower panel) begins to emerge.  A bubble representation highlights this better than color depth, with the disconnect in Africa being most evident (inverse pattern from South America):

In the highest tier, Lebanon, Ukraine, Azerbaijan and Paraguay show the highest penetration … and it is very high.  The approximate equivalent* of more than better than 1 in 100 of these countries’ poorest people have been served by Kiva microloans.  Mongolia and a host of Central and South American countries come next (Nicaragua, Costa Rica, Ecuador, Peru, Bolivia) within equivalent penetration rates of 1-in-150 to 1-in-250.   In order, Armenia, El Salvador, Tajikistan, Cambodia, the Dominican Republic, Togo, Kyrgyzstan, Chile and Moldova round out uppermost tier, up to 1-in-1000 equivalent penetration.

Back to visualization – the ManyEyes shaded map is still less than satisfactory to me.   Rearranging the data as a treemap may communicate better; in the graphic below, countries are binned by region, the boxes are scaled by the absolute size of the <$2/day population, and the color intensity shows the total amount of money loaned through Kiva.

Personally, I think the treemap approach does a better job at communicating relative need and gaps.

Of course for lending “decision support”, I’m more interested in even more finely tuned assessment of need.  Comparing poverty rates (x-axis) with lending penetration (y-axis) and throwing in a bubble scaling effect to denote absolute magnitude, brings out some patterns.

This is helpful for understanding (many countries with high Kiva “penetration” have comparatively small absolute populations in poverty) but less useful for quick and dirty decision support when it comes time to make loans.    So instead, I converted both the absolute poverty incidence (blue below, from the $2/day data) and the Kiva penetration data (or rather, its inverse; green below), percentile ranked all the countries, and combined the two into one “need ranking”:

In practice, I’ll be comparing this with my MFI rankings from last November.   Part of the reason some countries are underserved is due to shakier partner institutions.

Finally, by binning the “need ranking” into high, medium and low categories, and returning to trusty old Google Maps, I can provide a (perhaps) clearer picture of Kiva’s penetration worldwide:

Simplistic, to be sure, but hopefully the analysis and visualization drive home the need for continued penetration of microfinance (whether Kiva or otherwise) into African and Southeast Asian markets.

My final caveat is that “need”, of course, continues to exist worldwide.   Personally, I will continue to lend to “green” countries above, as they have mature microfinance networks and, perhaps, “manageably” sized poorer populations, where microlending can make a dent.   However, I will also now go out of my way to look for reasonably decent MFI partners (based on my earlier rankings) in “red” countries, where the need is demonstrably greater.

Technical note: To estimate “equivalent penetration” statistics, i.e., “1 in 100 poor people served”, I have taken the inverse of (total dollar amount loaned to a country, divided by the estimated size of the population living on <$2/day, divided by the average size of a Kiva loan [$749]).   I.e., I estimate total loans per poor-capita, and take the inverse to get “1 in x” statistics.   I call this “equivalent” since actual loan sizes vary, since loan recipients are often repeat customers, once they pay back their initial loans, and since recipients don’t necessarily originate from the lowest tier of wage owners ($2/day is a proxy for overall microloan need).   Even with the uncertainties, I think it really highlights the power of Kiva.  The fact that countries that develop strong MFI partner networks can reach “1 in hundreds“, not “1 in millions“, type penetration, is both surprising and inspirational.

Several months ago I reviewed Business Model Generation, a handy book and toolbox for preliminary sketching and development of, well, business models.   I was particularly intrigued by the book as it is simple enough to help untangle “business 101′s” for my customer base of engineers, scientists, and assorted government-side bureaucrats (of which I am one).   The author of this book has recently released version 1.0 of the accompanying iPad app, so … here’s the review. The short form – if you’ve already been converted to the virtues and value of the “business model canvas” (at left), and use it regularly, Version 1.0 of the toolbox is probably worth the $29.99 ante-up.

If you’re dabbling and think it might be useful (perhaps after reading the 72 page free short summary of the book available on the website), I’d still recommend it, though with the caveat that you should set your expectations based on a version 1.0 software release.

Before digging in, a couple of quibbles to go along with the release.  Both center around the fact that the iPad app release actually makes it almost harder to advocate for greater pickup and use of the Canvas tool.

1. The price – $29.99 is OK for the already-converted, power users, but it begs the question of why a “lite version” isn’t on the street as well.   Just in using the paper canvases, I have found many people who would probably be motivated to spend a few dollars to experiment with the canvas – but not $30.  The authors may be missing a customer segment.   I hope they consider a Light version or lower app costs once they recover their development costs.
2. The free downloadable canvas temporarily disappeared from the BMG website – but has reappeared now on wikipedia - so strike that quibble!   Huge props to the authors and developers for continuing to make the Canvas available with a Creative Commons license.

On to the review:

The iPad app interface is reasonably user friendly, and certainly supportive of individual, offline experimentation with business models.   However, as a substitute for printing a wall sized version of the canvas and working collaboratively through models, it is “not yet there”.   Even with an iPad2 and video mirroring to a projector, I just don’t see the app as an equal substitute – yet.   So my recommendation is to treat it as an individual / personal tool.   The experience is about as claustrophobic as one would expect taking an exercise intended for wall-size collaboration, down to iPad-size screen resolution, would be.   That’s not necessarily bad, just set your expectations right.

Making up for the interface limitations is the inclusion of a very nice underlying cost/revenue model.   The model is heavily biased towards “unit sales” type models, rather than services or levels of effort, but it can be retrofitted.   Extremely nice features include binning revenues according to different customer segments, value propositions, etc.   It appears that future versions will allow costs to be binned and sorted similarly, but that feature doesn’t seem to be available yet.    Adding an actual rubber-to-the-road calculation engine underneath the canvas itself is a real value add and worth paying for.   The interface to access the model isn’t seamless, but it is clever.

Critical (but incremental) features which would spice up a version 2 would include:

1. The ability to filter by assigned colors, to focus attention on individual value props or customer segments
2. PDF export
3. A native file format, so that canvases can be shared between users
4. Adding in the (latent?) feature to sort costs, not just revenues, by customers or value props
5. Adding an “annotation and markup layer” (arrows, highlight boxes, etc) to help show connectivity within models, as is done in the book

Another lifelogging experiment with GraphViz-based network mapping, this time on my library of books (sample size 356, managed with the excellent little Mac app Delicious Library 2, notable for its iSight-barcode-scanning magic, and no relation to the social bookmarking service).

In this experiment I map the Library of Congress classification system through to its Dewey Decimal counterpart.  (For those playing along at home, Delicious Library adds DD numbers automatically upon import.  Sadly, though, LOC codes must be searched and added manually).   The classification bins in the network diagram below are scaled based on the number of books in each bin (width, height, and font size scaled by linear factors with the square root of book count), and the edge connections are similarly weighted by book count.

Pause for rationale:   Why?   Answer:  Why not?

Library of Congress codes are on the left here, Dewey on the right.   A couple of things jump out at me: for purposes of coarsely bucketing my “top-level” interests, the much-maligned Dewey system seems to do a better job (fewer buckets, better categories), although at the expense of granularity.

More interesting is that the process of network mapping (allowing the graph layout algorithm to rearrange both top and second layer categories, driven by the cross-system edge connections) yields to me what seems a more seamless “sequential” ordering of each top-level system.   For example, both systems nominally list Technology (mostly engineering) after Science, but for me, the societal element of technology/engineering clearly places it closer to the social science domain (or is this just my ex-civil-engineer’s sensibilities in action?).   The network-rearranged system captures this.

Network mapping also “repairs” some Dewey defects, moving History & Geography back up to a “proper” (LOC-driven) place in between Philosophy and Social Science.

Interestingly, the mapping places Literature and Language close to, and “after”, Science, rather than buried deeper in the social sciences.   Topologically I might envision an even better classification to be circular rather than linear in nature, with Language “wrapping around” to connect back up with Philosophy or Religion.

A similar exercise with a larger sample of books (unbiased by personal preference selection) might yield a very interesting “meta-classification” system (not that the world necessarily needs one!).

A more readable PDF version is available.

Very little can be said to have been “good” about the brutal Alabama tornado outbreak of 27 April 2011.  I was fortunate enough to come through unscathed both in body and property, inconvenienced only by five days without power.  In a “glass half full” sort of way, five powerless days and public entreaties to stay out of the way of recovery efforts did have one upside: I got to catch up on quite a bit of reading.  Below are some drive-by reviews of an odd assortment of pent up reading material.

The Man Who Ran the Moon: James E. Webb, NASA, and the Secret History of Project Apollo

4 of 5 stars

Readers will find little new in “The Man Who Ran the Moon” (nothing in the way of a “secret history”), although its focus on an often-overlooked but central figure in the Apollo program – NASA Administrator James Webb – is a welcome supplement to the popular histories of the era. Webb’s views on the organization and management of the newborn agency are almost as interesting as his political dealings through the 1960s. The complex interplay between aerospace contractors and the Federal government is also given more exposure than conventional in popular texts.

A quick and easy read, “The Man Who Ran the Moon” is a worthwhile diversion for anyone interested in the history of NASA as an organization or the Apollo program itself, as well as public administration in general and Cold War-era beliefs about technocracies and their role in society.

Four Fish: The Future of the Last Wild Food

4 of 5 stars

I received this book as a gift from my brother, a far greater “afishionado” (groan) of all things finned than I. I’ll confess it sat idle for a while in the “medium priority” layer of my stack of reading material. It should have been higher.

Fitting comfortably in the “microhistory of a natural resource” new genre of books, “Four Fish” succeeds by drawing together and interweaving compelling personal, historical, economic and ecological narratives. Greenberg tackles difficult and tangled questions of sustainability of the oceans, but leaves readers with enough of a clear narrative and grasp of the issues to begin to form their own, informed, opinions. (This is no easy feat, and one that would have been fumbled by less competent authors).

One of the most interesting (and unexpected) turns is the treatment of aquaculture – domestication of wild species of fish – and its long term viability. The coverage is both balanced and deep. Greenberg also raises uncomfortable but compelling points about the disconnect between Western consumers’ faith in the power of the market, vs the realities of firm policy.

Even if you are only glancingly interested in the topic of “things with fins” (as I was) – this book is worth a try.

The Silk Road: Two Thousand Years in the Heart of Asia

3 of 5 stars

My interest in central Asia has been piqued since a recent homestay trip in Mongolia. Since then, the “-istan’s”, for me hidden behind the opacity of the Cold War for most of my life, have been a source of mild intrigue. Reading “The Silk Road” has added significant color – if not necessarily clarity – to my familiarity with the tangled history of the region.

The book is readable, if not necessarily fully accessible, to non-academics, presuming significant familiarity with the region and its peoples (both past and present). While not “lavishly” illustrated, it certainly draws from an eclectic sample of material which keeps the interest level up. Descriptions and details are rich. The historical narrative is far from linear, and overall the book would have benefitted from a greater use of maps than the single, stylized map in the front matter.   Quite honestly, if it had not be for the power outage, it would have taken me many more weeks to plod through this text, as I had struggled for three weeks’ worth of nightly reading to make it only halfway through.

I wouldn’t recommend this book as a casual read, although for fans of history – specifically multicultural and oft-overlooked corners of world history – it may be worth the plunge.

This is a “three birds with one stone” entry:

1. I wanted a test project to learn GraphViz and its related programming syntax for generating network diagrams.
2. I wanted to see how much semantic structure was built into my Delicious bookmarks, after two years and ~1190 entries worth of use.
3. I wanted to do the latter before Delicious gets transferred to its new owner and potentially gets dorked up.   (Yahoo’s utter apathy and neglect has at the very least been benign.  Although I am optimistic about AVOS…)

More specifically on #2, has enough information content found its way into my assigned tags to self-organize related keywords?  (I’m fascinated with the concept of folksonomies, even if a folksonomy technically requires more than one contributor.   

Garden-variety Wordle’s are easy enough to construct from Delicious bookmark tags, but are little more than toys, failing to reveal any linkage or semantic content.   (Other than the, I’m sure, accidentally fortuitous for me, choice of a Space Shuttle-type envelope for the “horizontal and alphabetical” tag ordering).   This makes for a nice bumper sticker, but not much else:

GraphViz (which happens to also have a very nice and convenient MacOS X interface) helps add back connectivity between tags.  Lots of connectivity:

The graphic is easier to penetrate in its PDF version.  In the diagram, I’ve connected pairs of tags that occur together in three or more bookmarks.  Tag bubbles are scaled (manually) by their overall frequency of occurrence.   The rest of the organization is done automatically by GraphViz in attempting to minimize the complexity (!) of the network’s edge (connection) layout.

Overall I’m neither disappointed nor ecstatic with the results.   There are nicely laid out “regions” on the diagram:  an “aerospace” zone in the middle-left, a “government” sector at lower-left, a “business” area at center-bottom, even a “shopping” nook at far-bottom.   Interestingly, the “travel” quarter got pulled up into the “aerospace and policy” regions, presumably due to multiple connections for  ”Europe” and “China” (which I have both travelled to, and also bookmark often from a space policy perspective).

Alternatively, there are limits to how much a complex, multidimensional semantic space can be “flattened” to 2-D.   The almost-lowest tier of tags sometimes has the appearance of a catch-all, and some tags have found themselves stranded there.

For a great discussion of folksonomies, check out Moritz Stefaner’s thesis - several years old but still an excellent read.

Below is a subset of the GraphViz code used to generate the network diagram.   It’s probably “kindergarten” in terms of GraphViz grammar level and any tips on improving the layout would be welcome.  Unfortunately, not much of this process was automated – scraping of the delicious bookmarks, scaling, and transcription to .gv format were a manual (although not painful) process.

digraph Delicious {
	graph [splines=true,overlap=false,concentrate=true,style="bold"]
	node [style=filled fillcolor=red]	

	"space" [shape=ellipse,regular=false,style=filled,fillcolor=green,height=24.5,width=73.5,fontsize=703.5];	
	"visualization" [shape=ellipse,regular=false,style=filled,fillcolor=green,height=22.36,width=67.1,fontsize=670.8];	
	"NASA" [shape=ellipse,regular=false,style=filled,fillcolor=green,height=20.82,width=62.45,fontsize=624.5];
	"travel" [shape=ellipse,regular=false,style=filled,fillcolor=green,height=16.832,width=50.4975,fontsize=504.975] ;

...
	"RP" [shape=ellipse,regular=false,style=filled,fillcolor=green,height=2.357,width=7.071,fontsize=70.71] ;
	"upper_stage" [shape=ellipse,regular=false,style=filled,fillcolor=green,height=2.357,width=7.071,fontsize=70.71] ;
	"vehicle" [shape=ellipse,regular=false,style=filled,fillcolor=green,height=2.357,width=7.071,fontsize=70.71] ;...
	space -> NASA [dir=none, weight=67];
	space -> launch [dir=none, weight=35];
	space -> policy [dir=none, weight=26];
	space -> science [dir=none, weight=26];

...
	booster -> reusable [dir=none, weight=2];
	booster -> RP [dir=none, weight=2];
	HLV -> Senate [dir=none, weight=2];
	HLV -> heavylift [dir=none, weight=2];

}

This will be a difficult entry to pen without being accused of being “hostile” to the emergent U.S. commercial space industry (which I am not).  It is – as with most posts in this blog – a discussion of data, more specifically, data from forecasts.   Very optimistic forecasts.   If it makes it any easier to swallow, very optimistic government forecasts.

First, some background.   Each year since 1995, the FAA, which has responsibilities for regulation of commercial space launches, has issued 10-year forecasts of future launch vehicle demand for both geostationary/geosynchronous (GSO, typically communications) and non-geostationary (NGSO, typically low earth orbit) payloads.   These forecasts form the backbone of this analysis.

Second, some context.   The commercial space bubble of the 1990s, fueled by high expectations in the communications satellite arena, is well-known.   Those responsible for launch forecasts were also well aware of systematic optimism in their integration and run-out techniques and have worked systematically to reduce them.   My goal here is not to savage the hard-working folks at FAA, but to highlight what happens when models quietly diverge from reality, and to help create better educated “consumers” of the official forecasts.

This graphic comes from the 2010 FAA/COMSTAC forecast, and demonstrates FAA’s good faith in providing both forecast and actual data against which to validate model skill.   Included in this chart are total (GSO+NGSO) launch demand forecasts from 2000 forward only.   An appendix in the 2010 report also includes NGSO forecast vs actual data for 2004 forward only.   The systematic positive bias is plain to see, and has typically been remarked upon within the reports themselves.

However, this chart has some serious defects which very much get in the way of communicating the wide disparity between available forecasts in any given year, and “truth”.   They are almost certainly unintentional defects, but the net effect is significant:

1. Forecasts prior to 2000 have been eliminated from this chart, and prior to 2004 from the NGSO-only chart.   This has the effect of removing some of the most egregiously bad bubble-era forecasts from the composite (see below for the effects of the actual bubble-era data).   In fairness, the forecast model has not been stationary over that time frame, undergoing tweaking and evolution.   Also in fairness, the caption represents this graphic as “Commercial Space Transportation Forecasts” – i.e., the performance of the FAA office – not, “Forecast model version 3.x forecasts”.   The implication is that this is a graphic of COMSTAC’s historical performance, and as such, it should include all data.
2. In compiling this graph, FAA has excluded all of the “robust market” alternative scenarios included in old forecasts (a practice stopped in the mid-2000′s when it became obvious that there wasn’t a robust market in the immediate future).  By instead showing only the conservative forecast, it obscures the fact that consumers of the forecasts in the Y2K time frame had a very wide range (one could argue, uselessly wide) of forecasts to “choose from”.   Again, see below for the forecast data.
3. Visual sleight of hand with the 2010 forecast gives the illusion of greater skill than actually exists.   Up until 2009, all forecasts are represented on the chart by lines, and “actuals” by the blue sand chart wedge.   However, the 2010 forecast is shown by seamlessly extending the actuals sand chart, not by including an additional line.    The effect is that all of the old 2010-2020 forecast run-outs now appear to live “within” an actuals wedge, the grey wedge (which of course, is nothing of the sort, it’s just another forecast).   This hides the fact that prior to 2009, the forecasts almost never intersected with reality, instead, systematically overestimating demand.   This is either intentionally deceptive (unlikely) or really unfortunately bad chartsmanship (likely).
4. By visualizing the joint NGSO and GSO forecast, the graphic downplays the severe errors in the very badly modeled NGSO forecast, since almost all of the “skill” in the model is from the comparatively stable GSO market.

Again, given FAA’s transparency and ongoing efforts to improve their models, this is almost certainly due more to poor chartsmanship and the difficulty of tracking and integrating heterogeneous data over multiple decades, than any intent to deceive.   To get a clearer picture, I’ve created the graphic at right, which shows instead the probability distribution of FAA forecasts for each year, compared against the actuals.  The intent is to show the full range of forecasts “available to choose from” for potential consumers of the data.   In compiling the forecasts, I have treated the “baseline” and “robust” market historical forecasts as equally valid, since in reality, consumers of the historical forecasts had little basis to select between them.

In this graphic, the dashed line shows the median of all forecasts, the innermost dark grey band contains the “most central 33%” of forecasts (33%-67% percentiles), the next lightest grey band contains the “most central 50%” of forecasts (25%-75% percentiles), and the outer very light bands contain 90% and 100% of all forecasts, respectively.  (In short, the intensity of grey shading is intended to cue viewers towards the most frequently occurring forecast values).   This view provides a clearer view of what a “bubble” looks like in practice.   The compilation is troubling not only for revealing the excessive optimism, but also for calling into question the model’s skill in predicting any trends whatsoever.  Other than the initial rise from 1996-1997 (forecasted two years too early), there is little evidence that the forecasts as a whole actually capture any signal in the actual time series.   A simple flat runout of the previous year’s actual launch rate (known as a “persistence” forecast in weather circles) would likely have provided better overall skill as a model (I need to verify this, but eyeballing the graphic, I’m 99% confident of that outcome).

The situation gets worse if we split the forecast into its non-geostationary and geostationary components, below.   The bulk of the forecast “skill” is coming from the stable GSO market (at bottom), although even there the forecasts are optimistically biased, lying atop an essentially flat trend.   The NGSO forecast performance contains even less “skill” than the aggregate above.

So – why share this bad news?   The point most certainly isn’t to be cute and condemn the folks at FAA who work hard to forecast an extremely fragile and unstable emergent market – a thankless and “no-win” task, almost by definition.  Nor is it to make life any more difficult for the struggling commercial space sector.

The point is to help create educated consumers of the official data, and to heighten awareness of how critically dependent we are on the industry’s own self-reported future manifests, which are almost certainly optimistically skewed (since they do not allow for unexpected contingencies).    Indeed, in the aggregate, the 10-year forecasts actually have performed worse in the first three years of each forecast, than in the last seven years – one hypothesis could be that this is bias by overly optimistic near-term industry manifests.

FAA continues to work to improve its model and has been fully open in communicating its performance.   Until, however, the forecasts are consistently able to eliminate positive bias, and to demonstrate an ability to actually detect trend signals, at minimum a deep “discount factor” should be applied before using them for policy or commercial purposes.

NASA’s ten field Centers (well, nine and the Jet Propulsion Laboratory, technically a Federally Funded Research and Development Center) do the “heavy lifting” of NASA’s missions.   Each Center has a portfolio of skills, capabilities and missions entrusted to their care.   Each Center also has its own “flavor”, focusing on those things it does best.   (Actually, this is as much culture as it is flavor).   I was curious to see how strongly each Center’s ”branding” might show up in our “official” tweets.

The map below doesn’t contain any major surprises, but there are some interesting details.   The very well known Centers (JPL, Johnson and Kennedy) make comparatively less use of their name and acronym when tweeting.   In contrast, my own home center of Marshall, as well as Dryden in the California desert, seem to blare out “please don’t forget us” by insistently using and re-using our names.

Also evident is how some of the outreach / public interest activities at some centers (the Great Moonbuggy Race here at Marshall, as well as our meteor spotting network, and the TEDxNASA event at Langley) can come to dominate the tweetstreams.   This isn’t a bad thing – these are great educational and outreach events in their own right, and reflect good pairing of social media with publicly engaging topics – but Marshall’s core capabilities and mission, including the primary ways it delivers value to the American public, have little to do with student competitions and meteor showers.   It underscores the challenges of maintaining clear and coherent “corporate branding” transmitted through an increasingly diverse array of media vehicles, even more so when by a public sector organization.

A higher quality PDF version is available here:  YouAreWhatYouTweet .

Note that this isn’t the only tweeting by NASA; many NASA missions tweet on their own.   The tweets shown here are only those by NASA’s “stewards”, or field Centers.

As technical notes, in creating this graphic I used the visualization tool at Neoformix to search and extract tweet histories, filtered out the common words “space” and “NASA”, and limited scope to mission-related nouns (i.e., verbs, adjectives and generic words like “video” or “images” were removed).  For the most part, I consolidated word stubs / roots where appropriate in scaling the words to match frequency.

Also, I’ve used absolute word counts (not relative) and applied a uniform cutoff across all Centers.   This has the effect of making our friends at Glenn look a little anemic, an unfortunate side effect of being a little bit newer to the Twitter game.   When I update to version 2, they will likely have more content “above the cutoff level”.

As a bonus for my NASA compadres at other field Centers:

How quickly after a scientist stops publishing does he or she fade to obscurity?   Well, that’s a broader question than I plan to answer here.   But I can do a little lifelogging data analysis to figure out if my expiration date has yet passed.

I switched careers from science to management in 2005, publishing my last journal paper in that year.   From 1993-2005 I published 11 journal papers as a first author, and participated as a co-author in 17 more.   In theory, that should leave a tidy trail of followup citations to determine if I’m fading from view.   Unfortunately, it looks like I’ll have to wait a little longer to find out.

The problem now is that the shelf life appears to be at least longer than the five years since I stopped publishing.   If I had continued writing (and collaborating), I would have expected the curves at right to rise much more than linearly (not exponentially, but more than linearly), due to a “composite citation” effect.   Instead, the curves are roughly linear, and haven’t yet begun to flatten out.

As an aside, interlinking of scientific journal citations across the internet is still in a fairly sorry and byzantine state.   There are a number of reasons, including that nonprofit professional societies that publish journals typically lack the resources to integrate across platforms with each other, or backfill legacy data, while commercial journal publishers lack the financial incentive.   All of which goes to say – it wasn’t easy to cobble the data above together across multiple earth science journal publishers (AMS, AGU, Elsevier, others), and before 2002, the data are pretty spotty.

Anyway – the impacts of disengaging from the scientific community are a little easier to see by looking at citation-per-year rates.    (At a first author citation rate of about 30 per year I certainly wasn’t on track to shake the foundations of the scientific establishment!)  The differences between first and co-authorship trends is instructive.

After 2005, a time when I made a pretty clean break with both publishing and the research community, the citation rate of my first author papers flatlined.  Eventually, the inevitable will happen and that trend will start declining … from which I’ll then be able to extrapolate my EYO (Estimated Year of Obscurity, of interest to absolutely nobody but me.)   It would certainly be nice if this were greater than my remaining life expectancy, but I’m not optimistic.   Then again, I didn’t get into earth science to jump on the fast track to immortality.

In contrast, the citation rates for papers I participated as a co-author on continues to rise.   That suggests one (or both) of two things:

1. The colleagues who were leads on these papers were systematically better at conducting and publishing more relevant and enduring research (certainly possible, although if I look at the citation rates paper by paper, it doesn’t seem to match the data).
2. The colleagues who were leads continued to engage, continually increasing their own visibility within the community as well as that of their publications, through self-citation, leading to even  greater awareness of publications and increasing citation rates.

Note that I’m only tracking citations for the controlled set of papers I was author or co-author on, not citations for all authors I participated with.   The short form is: my colleagues’ stock continues to rise, while mine has – as expected – stalled out.

To readers who have slogged this far, my thanks for enduring what, in retrospect, is an entry high in narcissism and low in protein content.