1 of 29

Sudden Attention Shifts on Wikipedia During the COVID-19 Crisis

Manoel Horta Ribeiro*, Kristina Gligorić*, Maxime Peyrard*, Florian Lemmerich, Markus Strohmaier, Robert West

* = contributed equally

2 of 29

The impact of COVID-19 transcends health-related issues

(And thus, research ought to identify how the pandemic has impacted human needs, interests and concerns)

3 of 29

Challenges:

Immediacy
Scale

(Salganik 2019)

4 of 29

Wikipedia data is...

… “always on!”
... global!
… freely available!

In our goal of identifying how the pandemic has impacted human needs, interests, and concerns
We turn to wikipedia
Unlike most previous events that directly impacted so many lives around the world, the COVID-19 pandemic developed in a time of widespread Internet access.
Wikipedia, the world’s largest encyclopedia and one of the most visited sites on the Web, captures rich digital traces from readers
Allows us to do so in a systematic way due to its nature:

immediacy: “always-on” nature of usage logs allows us to analyze information seeking patterns at any time
At scale: Globally and freely available

Bob feedback:

A bit opportunistic! Maybe change a bit these slides

Highlight specific articles: Toilet paper article!

Systematic way of doing

There is uncertainty! Coronavirus!

What about the rest of wikipedia?

Random movie!!

5 of 29

6 of 29

Idea

Use Wikipedia pageview data to understand how information seeking behavior changed during the pandemic!

7 of 29

We analyze Wikipedia access logs in 12 wikipedia language editions

We selected 12 languages that have Wikipedia editions of various sizes. When choosing language editions, we considered (1) the size of the edition, (2) whether the language was spoken in relatively few countries, (3) the kinds of non-pharmaceutical interventions imposed in those countries. The overarching goal in selecting the languages was to have relevant language editions representing different attitudes towards the crisis, preferably from languages easily tied to one country. For each language edition, we obtained publicly available pageview statistics,1 which specify how frequently each article was accessed daily between 1 January 2018 and 31 July 2020, aggregating both the desktop and the mobile versions of the site.

8 of 29

🇮🇹 Italian Wikipedia

9 of 29

🇮🇹

10 of 29

🇮🇹

11 of 29

🇮🇹

12 of 29

🇮🇹

13 of 29

🇮🇹

14 of 29

(Angrist and Imbens 1995)

Difference-in-Differences

15 of 29

(Angrist and Imbens 1995)

Difference-in-Differences

16 of 29

(Angrist and Imbens 1995)

Difference-in-Differences

17 of 29

Difference-in-Differences

(Angrist and Imbens 1995)

18 of 29

Difference-in-Differences

(Angrist and Imbens 1995)

19 of 29

Difference-in-Differences

(Angrist and Imbens 1995)

20 of 29

Difference-in-Differences

(Angrist and Imbens 1995)

21 of 29

Difference-in-Differences

(Angrist and Imbens 1995)

Difference-

in-Differences

22 of 29

23 of 29

24 of 29

25 of 29

26 of 29

27 of 29

Image by EpochFail, licensed under CC BY-SA 4.0

ORES: Automatic categorization of Wikipedia pages

28 of 29

29 of 29

Wikipedia seems to be important in the time of crisis
What got increased permanently are not usual “needs”: Third, our results suggest that most of the topics that experienced persistent increases (e.g., VIDEO GAMES and INTERNET CULTURE) are not related to needs often framed as “basic”, such as such as safety and physiological needs".