Ranking the Pantheon

A previous post described the MIT Pantheon, another project which used Wikipedia data to rank historical figures.   We (meaning Charles, of course) extracted their rankings and matched them to our historical significance rankings, so we could compare them.   There is some subtlety in algorithmic name matching, such as determining whether our "Jesus" is the same person as their "Jesus Christ", but we succeeded in matching 10,116 of the Pantheon names to our Who's Bigger rankings.  This is roughly 90% of the total, providing a reasonable basis for comparison.

First off: it is clear that there is substantial agreement among our placement of historical figures, with a Spearman rank correlation of 0.65 between us and them.   Both sets of rankings incorporate aging as part of the methodology, so much of this agreement rests on our preferences for the tried and true.  The Who's Bigger rankings of these figures have a rank correlation of 0.58 with year-of-birth (older historical figures being more highly ranked), while the comparable number is 0.53 with Pantheon.

More revealing is to look at the extremes: the figures whom we assign very different ranks from them.   In particular, we computed the difference between our ranks (Pantheon - us) and present the figures with the largest and smallest differences.   This is not a perfect statistic, since Pantheon ranks less than 12,000 people while our numbers go well above 800,000.   But it is revealing none the less.

Diff       Panrank  BigRank       Name                                    Who's Dat?
10120     10521       401        'John Marshall'                            Chief Justice of the US Supreme Court
10058     11184     1126        'Donald Bradman'                       Great Cricket champion
10027     10823       796        'William H. Seward'                    U.S. Secretary of State (bought Alaska)
9963       11077     1114        'Gough Whitlam'                         Australian Prime Minister
9933       10812       879        'John Churchill 1st Duke of Marlborough'     English Statesman
9915       10802       887        'George Washington Carver'       African-American Inventor
9886       10405       519        'Tipu Sultan'                                Ruler of the Kingdom of Mysore
9735       10146       411        'John Jay'                                     Early U.S. Statesman
9536        9935        399         'John C. Calhoun'                       U.S. Senator /VP (nullification)
9454        9886        432         'Susan B. Anthony'                     U.S. Suffragist (women's right to vote)
9439      11243      1804         'Alexander Mackenzie'                Second Prime Minister of Canada
9243      10064        821         'Abigail Adams'                           Wife of President John Adams
9215      10729      1514         'Robert Menzies'                          Longest serving Australian Prime Min.
9207      10917      1710         'Robert Byrd'                               Long-serving U.S. Senator
9175      10406      1231         'Sojourner Truth'                         African-American abolitionist
9171      10562      1391         'Lucille Ball'                                TV Comedian (I Love Lucy)
9171        9330        159         'John A. Macdonald'                   First Prime Minister of Canada
9165      10466      1301         'Edmund Barton'                         First Prime Minister of Australia
9130      10318      1188         'Mary Todd Lincoln'                   Wife of President Abraham Lincoln
9008      10086      1078         'Svetlana Kuznetsova'                 Russian tennis star

Almost all of these figures are from the English-speaking world: United States, Canada, Australia, Great Britain.   It is no surprise that our methods (which only analyze the English language Wikipedia) generally rank these people higher than Pantheon (which analyzes editions from all languages).  I personally recognize 14 of the twenty names here, and think they are generally quite Big, although I cringe a bit where some of our rankings are clearly too high (particularly Sultan and Kuznetsova).

The major American figures here are generally from the 19th century, which makes sense given the difference between our aging model and the one employed in Pantheon (full disclosure: Pantheon has recently changed its rankings, and what we have here may not be their current rankings).   In particular, our rankings have fully discounted a historical figure 160 years after birth, while they continued historical discounting arbitrarily far into the past).   Thus 19th century figures have generally achieved steady state by our analysis, so we value them relatively higher than Pantheon would.

The other side of the coin are the people who Pantheon ranks very much higher than we do.   The figures below all ranked in the bottom half of Wikipedia figures by our analysis, yet were identified by Pantheon among the 12,000 most interesting figures for analysis:

Diff            Panrank  BigRank       Name                                    Who's Dat?

-472241      8052         480293  'Alexandra Stan'                       Romanian singer and model

-484757      11086      495843  'Serge Haroche'                         French Nobel Prize winner in Physics, 2012

-493874      9471        503345  'Lola Pagnani'                            Italian actress

-495688      11148      506836  'Stephane Lannoy'                    French soccer referee

-497360      10133      507493  'Olivier Giroud'                         French soccer player

-517354      11160      528514  'Wouter Weylandt'                   Belgian professional cyclist killed in 2011

-525449      9576        535025  'Nathalia Dill'                            Brazilian television actress

-525475      10601      536076  'Milos Zeman'                           Current president of the Czech Republic

-525633      11232      536865  'David J. Wineland'                  Nobel Prize winner in Physics, 2012

-526148      10774      536922  'Gianluca Ramazzotti'             Italian singer-songwriter

-555909      11029      566938  'Linda Maria Baros'                 Contemporary French poet

-558970      10942      569912  'Jules A. Hoffmann'                  French Nobel Prize winner in Medicine, 2011

-573789      11144      584933  'Pastora Soler'                           Spanish Eurovision singer

-581161      11286      592447  'Sun Yang'                                Chinese Olympic swimmer

-601660      10310      611970  'Kevin Grobkreutz'                 German soccer player

-607491      11318      618809  'Missy Franklin'                      American Olympic Swimmer, 2012

-613223      11278      624501  'Brian Kobilka'                        American Nobel Prize winner in Medicine, 2011

-632278      11224      643502  'Lobsang Sangay'                    Prime minister in exile for Tibet

-685152      10556      695708  'Bernice Bejo'                          French-Argentine actress

-689256      11296      700552  'Vaclav Pilar'                            Czech soccer player

-693543       9577       703120  'Raphael Varane'                      French soccer player

-717448       11231     728679  'Ludmilla Radchenko'            Russian model and active 

-751460       10907     762367  'Anton Lamazares'                   Contemporary Spanish painter

-803441       11270      814711  'Petr Jiracek'                             Czech soccer player

These people are generally Europeans, who have the easiest time rising to the Pantheon Wikipedia language threshold.   They are also all very contemporary figures, many of who achieved their greatest renown for achievements occurring after the Wikipedia edition we analyzed in our rankings (October 11, 2010), so presumably they would be ranked somewhat higher if we reran our analysis today.

However, I personally only recognized one name here, and it required some prompting. Bernice Bejo was the lead actress in "The Artist" which, by the way, was a wonderful picture.   These people would generally not be in my 12,000 most significant (or famous) historical figures, but Pantheon's objectives are somewhat different than ours.   My guess is the both groups are content with our ranking differences given our different motivations.

Big Data Done Wrong?

An Op-Ed piece in today's New York Times by Gary Marcus and Ernest Davis present Who's Bigger as the seventh of eight (or nine) problems with Big Data, specifically "giving scientific-sounding solutions to hopelessly imprecise questions".  They acknowledge that we get many things right, but complain about "egregious errors".

But guys: given a 379 page book with thousands of rankings to pick from, your killer example is that we ranked Francis Scott Key at position 19 on the poets list?   If they don't have a complaint until position 19 on one of several dozens of tables in our book, well, we must be doing pretty darn good.

But their chosen example is illuminating, because it gets to the heart of what our rankings are and are not designed to do.  Our book carefully claims to measure "historical significance" or "meme strength", not "importance" as they insist on misrepresenting in the article.

So how historically durable will the Francis Scott Key meme be, say 100 years from now?   If there is still a United States stuck with the same national anthem (I'd take that bet), then we can be pretty certain the Marcus and Davis great-great-great-grandchildren will learn Key's words and the story behind his work.

"Oh say can you see?"  Only if you are willing to look at what data is actually trying to tell you.

Time Magazine's College Rankings

Time Magazine has launched an interactive feature ranking colleges by the prominence of the Wikipedia pages of their living graduates.   Harvard appears to be the top dog by this measure, just edging past Stony Brook (which again failed to make its way into the NCAA basketball tournament, the event which inspired Time's feature):

Their ranking methodology includes certain Wikipedia variables analogous to what we have used, including length and links in/out of the page -- which serves as a poor man's version of PageRank.  But PageRank is much better for meaningful notions of importance: links into a page only matter if they are from prominent individuals, and links out have little obvious meaning except that it should be correlated strongly with article length.

The other aspect of such an analysis is properly attributing alumni to schools.   The Wikipedia categories give fairly unreliable annotations, although after checking I can confirm that Pat Benatar in fact did attend Stony Brook for a year before dropping out.  I guess we "hit her with our best shot".

The Pantheon

I was reading my Sunday New York Times when my heart skipped a beat.   There in the magazine was an article ``Who's More Famous than Jesus?'' which had to, just had to, be about our Who's Bigger rankings.

Well, it wasn't.  A project at MIT called Pantheon was the source of the article.   Pantheon also uses analysis of Wikipedia data to rank the fame of historical figures.    I will confess to a little sense of Schadenfreude in reading the comments complaining about theie rankings, including:

• Their bias towards Americans in particular and the Western world in general.
• That they contain too few women in highly ranked places
• Gleefully pointing out occasional mechanical misclassifications of individuals (particularly problematic was identifying John Wayne Gacy as a comedian instead of a serial killer)
• Making too big a deal of small differences between rankings of closely matched people
• Complaining that Wikipedia is not a reliable source to analyze world culture.

This all sounded very familiar, because these comments have been made about our rankings as well.

It seems worthwhile to compare our rankings and methodology with that underlying Pantheon.  There are several differences between our approaches to using Wikipedia as a resource:

• Languages -- Pantheon makes use of the multiplicity of Wikipedia language editions in its analysis.   To be ranked as truly famous one must appear in at least 25 different language editions.  This would make the rankings more inclusive of world opinion than our English-only analysis, although reader comments still complain about the Anglo-centric bias of the results.
• Variables -- Of the Wikipedia variables we employ in our rankings (two forms of PageRank, hits, edits, and article lengths), Pantheon only employs page hits.   Thus their notion of Fame is more akin to our notion of Celebrity (which loads heavily on hits).  Gravitas is the other component of historical significance, which we found loading most heavily on PageRank.   Thus we would expect their rankings to over-emphasize popular culture ahead of ours.
• Corrections for Time -- Pantheon employs an exponential decay model of fame in an attempt to correct for the recency bias of fame.   This overcompensates for the passage of time: six of the Pantheon top ten were ancient Greeks, with three others (Jesus, Confucius, and Julius Caesar) living 2,000 or more years ago.   The most recent member of the Pantheon top ten only gets us to the Renaissance (Leonardo da Vinci).  Our aging model is more sophisticated, and calibrated to appearances of names in 200 years of scanned books / Google Ngrams.
• Validation -- Their website includes an analysis of how their rankings compare to performance in three sports domains: Formula 1 racing, tennis, and swimming.  Our book discusses how our rankings compare to sports statistics (particularly with respect to baseball), but we also perform a more general set of validation tests, including correlations against 35 published rankings, prices of collectables including paintings and autographs, and public opinion polls.

To their credit, their website is fun to play with and features a host of interesting visualizations.

But how good are the rankings?  It is easy to cherry-pick any set of rankings for things that look weird. They name Rasmus Lerdorf (developer of the programming language PhP, who frankly I had never heard of) among their top 11,000 people, on the strength of being in more than 25 Wikipedia editions (he is actually in 31).  By comparison, we have him as the 51,670th most significant figure.  They rank Justin Bieber at 671 to our 8633, and Johnny Depp at 203 to our 2739, suggesting an over-emphasis of celebrity at the expense of gravitas.

But the right way to compare rankings is through validation measures.   This takes work, but I hope we can do such a study soon.  We will report our results here when we do.

First Ladies: Siena rank vs. Skiena rank?

As part of a collaboration with C-SPAN, Siena Research Institute has just presented the results of its latest historian poll ranking the top American First Ladies, i.e. the wives of the presidents.  They have conducted five such rankings over the past 31 years, through a process of asking experts where they rank in such categories as Background, Value to the Country, Leadership, Being her own Woman, Accomplishments, and Courage.

We constructed our own rankings of First Ladies in Who's Bigger, through Wikipedia analysis, so it is an interesting exercise to compare our rankings.  Bottom line -- we come off quite well.

We agree with the poll's selection of Eleanor Roosevelt as the top first lady.  In fact, six of our top ten appear among the top ten in the Siena Poll.   All of our top ten rank in the top half of the 38 first ladies ranked by Siena, except for one.   We regard Mary Todd Lincoln as the fifth most significant first lady, where they rank her as the 30th best.   There is no contradiction here: the meme associated with Mary Todd Lincoln is of a needy, crazy woman tormenting her husband when he really had other things to deal with.  She was indeed historically significant, but not in a favorable sense.

Our ranking of the ten least significant First Ladies included three Siena didn't bother to rank.  Chester Arthur and Martin Van Buren were widowers when they entered the White House, so it questionable whether we should have considered their spouses at all.   William Henry Harrison died after a month in office, barely leaving his wife with time to unpack.  Our remaining seven slots are filled with four from Siena's bottom ten (the wives of Taylor, Pierce, Fillmore, and McKinley), with the remaining three all ranking in the bottom half of the Siena poll.

These results demonstrate the ability of our ranking methods to tease apart significance even of relatively minor historical figures (the average first lady ranks in the neighborhood of 15-20,000 or so). My suspicion is that Wikipedia-based rankings does particularly well at this task because the expert panelists probably snuck peaks at the encyclopedia to help answer the poll!   I expect very few historians could keep straight the accomplishments of all the first ladies without a refresher.

A Moment of Wikipedia Glory!

Charles and I were surprised and flattered to discover that Who's Bigger has officially been granted its very own Wikipedia page, in English.  By repeating our computational analysis in the future, we will now be able to rigorously determine whether we are bigger than, say, the Bible.   OK, maybe this is somewhat aspirational, but as the first of my five books to earn its own Wikipedia page Who's Bigger already becomes my biggest book by default.

The coolest thing is that for eight hours starting 08:00, 21 January 2014 (UTC) our book held pride of place under Did you know? on the front page of Wikipedia!  Such placement matters.  Statistics show that our page has been accessed 7596 times over the past thirty days, exactly 5144 of which came on January 21.

Professor, R.I.P.

What one is exposed to as a youth can have a tremendous impact on future life paths.   I have spent my full working life as a college professor, but did not come from an academic family. During my youth, there was only one professor I was really aware of, and his model no doubt influenced my choice of career in ways that I am not fully aware of.

I feel moved to acknowledge the influence of Russell Johnson, the Professor on the TV show ``Gilligan's Island'', who passed away today.  He never seemed constrained by disciplinary boundaries; a generalist with deep knowledge of every subject, and a flair for creating high technology items out of coconut shells.  His model of the nerdy academic wandering his tropical island paradise was so compelling that maybe it helped turn me into a nerdy academic wandering my not-so-tropical, Long Island not-so-paradise...

This seems an appropriate opportunity to rank the seven stars of Gilligan's Island by their historical significance:

• Jim Backus (Mr. Howell) 15,374.
• Alan Hale, Jr. (The Captain) 23,841.
• Bob Denver (Gilligan) 26,025.
• Tina Louise (Ginger) 28,365.
• Dawn Wells (Mary Ann) 35,815.
• Natalie Schafer (Mrs. Howell) 43,321.
• Russell Johnson (The Professor) 48,384.

These actor rankings grossly reflect my sense of reality.   Jim Backus was a genuine movie star (remember him as the father in ``Rebel without a Cause''?) who achieved his greatest cultural role as the voice of Mr. Magoo.  I would have ranked Bob Denver ahead of Alan Hale, but the captain appeared in several movies prior to his role on the show.   None of the four supporting actors ever had any really significant roles outside the show, a typecasting fate which seemed to strike many of the television actors of the era.