C. Candia-Castro-Vallejos, Cristian Jara-Figueroa, César A. Hidalgo

Macro Connections - MIT Media Lab

Center for Research in Social Complexity


In 2016 hundreds of famous people died. David Bowie, Prince, Muhammad Ali, Fidel Castro, Carrie Fisher, Gene Wilder, George Michael, Vera Rubin, and Thomas Schelling, are just a few names on this list. But was 2016 a particularly bad year? Or will 2017 be "worse"?

Many people think that 2016 was particularly bad, but data tells us that it is not an exception. In fact, we should expect more famous people to die in 2017 than in 2016. Why? The answer is simple: because the number of famous people has increased over time.

Of course there are a few caveats you may be considering, such as: who qualifies as famous? Or, is the increasing number of famous people just a consequence of global population growth?

To answer the first question, here we use a simple definition of fame that we can implement using data. We define someone as famous if we can read about them in many languages. How many languages? 20 or more to be exact. That is, we focus on the 29,421 people who was present in more than 20 Wikipedia language editions as of February, 2016.

Of course this data has its limitations. Trust us, we have thought about those extensively. Nevertheless, the number of languages a biography has in Wikipedia is a scientifically validated, yet simple and imperfect measure of a person's fame or memorability (since Wikipedia is a form of cultural memory). Consider the singer David Bowie. In Wikipedia you can read about him in 104 different languages. How about the actor Gene Wilder? 84. And the economist Thomas Schelling? In 48. Certainly, this does not mean that Bowie's work was more, or less important, than that of Schelling. It simple means that more Wikipedians (and probably more people) are aware of Bowie's songs than of Schelling's theories (which is reasonable, given the global popularity of some of Bowie's songs).

The second question was whether the number of famous people has increased simply because the population of the world has increased. Our data shows that this is not the case. For centuries, the growth in famous people has been outpacing that of global population. As you can see in this paper and in this short talk, the number of famous people born in a given year used to be a fraction of global population prior to the invention of printing, and also, for the 200 years after printing (although it was a larger fraction). Since the late seventeenth century, however, the number of famous people born in each year has been proportional to the square of global population. That is, the number of births of famous people, divided by the population of the world at that time, has been increasing linearly over time. Moreover, that proportionality constant has increased with the introduction of new communication technologies. The slope that emerged with the popularization of shorter forms of printing, like journals and newspapers in the late seventeenth century, increased with the introduction of new communication technologies, like film, radio, and television. So in the twentieth century we produced famous people at a rate we never did before.

So now that we have cleared out these caveats, we can sink into the data and look at how many famous people we expect should die in 2017.

Figure 1 shows the number of people with a presence in more than 20 language editions of Wikipedia who have died each year since the year 2000. This number has increased linearly from 86 in the year 2000 to 195 in 2015. In 2016, actually, we observe that less famous people died than expected. So what may be causing our feeling of perceiving more deaths?

One answer may be that those who died this year were particularly famous. After all, you can read about David Bowie in more than 100 language editions of Wikipedia but in Figure 1 we are requiring only 20. So not all of these celebrity deaths are of Bowie's caliber. Figure 2 repeats the same exercise using higher thresholds: more than 20 languages, 35, 50, and 70. The last category (more than 70), is that of the mega-famous people, those who we are likely to hear about in the news and whose death is likely to echo in social media. Here 2016 looks exceptional. In 2016 we observed the death of 16 mega-famous (L>70), while in 2015, 2014, and 2013, this numbers were respectively 9, 10, and 14.

The list of the mega-famous who died this year, sorted by L is:

Name L Name L
1. Fidel Castro 135 9. Bhumibol Adulyadej 84
2. Muhammad Ali 124 10. Dario Fo 82
3. David Bowie 104 11. Leonard Cohen 80
4. Umberto Eco 92 12. Prince 76
5. Shimon Peres 86 13. Islam Karimov 76
6.Carrie Fisher 86 14. Nancy Reagan 76
7. Johan Cruyff 85 15. Imre Kertész 74
8. Gene Wilder 84 16. George Michael 72

So how has the median age, or year of birth, of the globally famous people who passed away each year changed over this period? Are we observing the death of increasingly older people? To some extent. On Figure 3 we observe, in the year 2000, the median birth year of the celebrities who died was 1920. Meaning 80 years old. In 2015 and 2016 it was, respectively, 1932 and 1930, or 83 and 86 years old. So the median age of the deceased individual has increased.

This increase, is important, however, because it we are now starting to see the death of people who did their best work in their 60s, 70s, and 80s, and whose fame was amplified through television.

Looking at the list of the sixteen mega-famous people who passed away this year helps us understand who these people were, and what contributions they did. This list includes a large number of performing actors, but also, political leaders, like Fidel Castro, Shimon Peres, or king Bhumibol Adulyadej. To explore the larger list of almost 200 people, however, we need to plot the occupational categories and the birthplaces of the people who passed away each year.

Figure 4 looks at the occupational categories of the famous people who passed away each year. You can click on the names of categories to turn these on or off. The most popular category is performing artists. Its representation has increased over time. In the year 2000, performing artists deaths represented 29 percent of all celebrity deaths. In 2016, they were 36 percent of the total. Scientists, on the other hand, have been almost constant as a proportion of total celebrity deaths. They represented 10.5 percent in the year 2000 while in 2016 they were 9.9 percent. Celebrities from humanities, have slightly decreased as a proportion of total celebrity deaths. They represented 15 percent in the year 2000 while in 2016 they were less than 11 percent.

Should we expect the number of famous people who die each year to continue to increase? Probably for the next years, but not forever. The rise of communication technologies in the last six centuries, from printing to social media, has increased the number of people who see their work amplified and remembered (although fame is fleeting, meaning that not everyone that is famous in a time period is remembered forever). Despite this rise in communication technologies, we may soon reach a time when what will limit the number of famous people we produce will no longer be our means of communication, but our limited attention and human memory. Maybe, we are already there.


About the Author
Cristian Candia-Castro Vallejos is a Chilean physicist pursuing his PhD degree in Social Complexity Science at CICS. He lives in Cambridge, Massachusetts where he works as research assistant at Macro Connections, MIT Media Lab. His interests are related to analytic studies on Collective Memory, Economic Complexity, Collective Learning, Big Data and Social Dynamics.

About the Author
Cristian Jara Figueroa is a Ph.D. student at the MIT Media Lab focusing on how groups of people accumulate knowledge and knowhow. He is in charge of Pantheon, an effort to discover patterns in history by focusing on biographical data.

About the Author
César A. Hidalgo leads the Macro Connections group at The MIT Media Lab and is also an Associate Professor of Media Arts and Sciences at MIT. Hidalgo's work focuses on collective learning. That is, the learning that takes place in teams, organizations, cities, and nations. In his lab he develops analytical tools to improve our understanding of how collective learning takes place, and also, he develops data visualization and analysis tools designed to improve the collective learning of organizations. Hidalgo's academic publications have been cited more than 8,000 times and his visualization engines have been viewed more than 100 million times. Hidalgo is the author of Why Information Grows (Basic Books, 2015), the co-author of The Atlas of Economic Complexity (MIT Press, 2014), and a co-founder of Datawheel LLC. He lives in Somerville Massachusetts with his wife Anna and their daughter Iris.


my widget for counting