Quantifying national information interests using the activity of Wikipedia editors
This document analyzes Wikipedia editing data to quantify national information interests and see if the globalization of technology has led to globalization of interests. The researchers analyzed over 23 million edits from 234 countries to see which countries edit the same Wikipedia topics. They developed a filtering model to identify statistically significant links between countries based on shared editorial interests. The results showed countries still have localized interests, as the network of shared interests grouped countries into clusters largely aligned with geographical and cultural boundaries, such as European countries editing about similar people, places, and events. Thus, the study found that despite globalized technology, people's information interests remain local.
Quantifying national information interests using the activity of Wikipedia editors
- 1. Quantifying national information interests using the activity of Wikipedia editors F. Karimi, L. Bohlin, A. Sammoilenko, M. Rosvall , A. Lacincinetti IceLab – Umeå University, Umeå, Sweden GESIS – Leibnitz institute for social science, Cologne, Germany
- 2. Background • Historically spreading of information was limited by the word of mouth, social gatherings, written documents, roads that connect cities etc. • Today advances in the electronic communication have revolutionized how we access and perceive information. • Many people around the planet have more equal access to the same types of information.
- 3. Does the globalization of technology enforce globalization of information interests?
- 4. • Information is being added and edited by the citizens of the world • No centralized authority • Widely used across different countries • Available in many languages • Contains all sorts of information Why Wikipedia?
- 6. • > 1 million random Wikipedia articles in all language editions including English • Retrieve the location of unregistered editors. • > 23 million edits from 234 Countries and 248 languages Method
- 7. Linking locations if they edit the same WP topics.
- 8. ?
- 9. Filtering model • Expected chance that two countries edit the same article with size n, follows a multinomial distribution. < AB >= n(n−1)pA pB var(AB)= n(n−1)pA pB((6−4n)pA pB +(n−2)(pA + pB )+1)
- 10. Extracting significant links z-score =Σ ((empirical weight – expected weight) / standard deviation) • Threshold: 5% pvalue with Bonferroni correction Significant weight = z-score - threshold
- 11. Network of information interests
- 12. Clusters of countries with similar information interests
- 13. Clusters of countries with similar information interests Austria Germany Belgium Netherlands SwitzerlandFrance Luxembourg Monaco Liechtenstein 18 Italy, San Marino, Holy See (Vatican City) aCaribbean Islands in the list are: Jamaica, Trinidad and Tobago, Saint Lucia, Barbad and the Grenadines, Belize, US Virgin Islands, Dominica, Cayman Islands, British Virg Supplementary Table 2 Top 10 Wikipedia articles co-edit based on the filtering analysis Rank DE-AT SE-NO 1 Christina Stürmer Tipuloid 2 Steffen Hofmann Dansban 3 Piefke Erik Ha 4 Klagenfurt Sweden 5 Kottan ermittelt Petter Jö 6 Der Bulle von Tölz Causerie 7 Puls 4 List of t 8 Austrian legislative election, 2006 Allmänn 9 Wolfgang Ambros Fredrik 10 Single cable distribution Daniel Ö
- 14. Middle East North America Russia & Eastern Europe South America Scandinavia Interests highways
- 15. Conclusion Despite the globalization of technology we still care about local information.