Most popular contents requested by users in different Wikipedia editions
Antonio J. Reinoso1, Juan Ortega-Valiente1, Roc´ ıo Mu˜ noz-Mansilla2and Carlos Le´ on1
1Department of ICT Engineering, UAX, Avda. de la Universidad, 1, Vva. de la Ca˜ nada, Spain
2Department of Computer Science and Automation, UNED, C/ Juan del Rosal, 16, Madrid, Spain
email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org
Wikipedia, Use Patterns, Traffic Characterization, Content Categorization
This paper aims to analyze how the most requested and contributed contents in Wikipedia may significantly
vary depending on the considered edition. The on-line Encyclopedia has become a prolific research topic,
mainly in aspects related to the assessment of its contents and in its evolution forecasting. However, very
little effort has been devoted to deal with the kind of use given to Wikipedia by its visitors, either occasional
or subscribers. Thus, our work aims to explore the utilization made of Wikipedia through a classification
of the most requested and contributed contents in some of its editions. This way, we will be in position of
determining which type of contents attracts the highest numbers of visits and contributions in these editions
and which can be a good indicator of the use given to them by their respective community of users. Apart
from the subsequent comparison purposes, such examination may reveal interesting topics such as the trans-
mission of tendencies over the different Wikipedia editions, as well as particular user patterns exhibited by the
corresponding communities of users.
Wikipedia continues to be an absolute success and
stands as the most relevant wiki-based platform. As a
free and on-line encyclopedia, it offers a rich collec-
tion of contents, provided in different media formats
and related to all the areas of knowledge. Undoubt-
edly, the Wikipedia phenomenon constitutes one of
the most remarkable milestones in the evolution of
encyclopedias. In addition, its supporting paradigm,
based in the application of collaborative and coopera-
tive efforts to the production of knowledge, has been
object of a great number of studies and examinations.
Wikipedia is organized in about 285
each corresponding to a different language. All these
editions add up more than 22 million articles2, which
correspond to encyclopedic entries about particular
subjects, events or people. Wikipedia articles not only
address topics from academic disciplines, such as sci-
entific or humanistic subjects, but also from areas re-
lated to music, sports, current events and so forth. Fo-
cusing on the audience, the overall set of Wikipedia
editions attracts approximately 15,000 million visits
1http://meta.wikimedia.org/wiki/List of Wikipedias
(Retrieved on 6 June 2012)
2http://meta.wikimedia.org/wiki/List of Wikipedias#-
Grand Total (Retrieved on 6 June 2012)
gument for the popularity gained by Wikipedia and
contributes to reinforce its massive acceptance by the
As a result of this relevance, Wikipedia has turned
into a subject of increasing interest for researchers4.
This way, quantitative examinations about its articles,
authors, visits and contributions have been performed
in different studies such as (Ortega et al., 2007) and
(Tony and Riedl, 2009). The quality and reliability of
the offered information has propitiated a prolific re-
search field where several techniques and approaches
have been conducted: (Korfiatis et al., 2006), (Giles,
2005) and (Chesney, 2006). In addition, Wikipedia’s
growth tendencies and general evolution have also
been largely addressed in studies such as (Capocci
et al., 2006) and (Suh et al., 2009). Several other
works have focused on particular aspects, such as mo-
tivation: (Kuznetsov, 2006) and (Nov, 2007), con-
sensus:(Kittur et al., 2007),
and (Vi´ egas et al., 2007) or vandalism: (Priedhorsky
etal., 2007). Bycontrast, fewstudies (Urdanetaetal.,
2007), (Reinosoetal., 2010)or (Reinoso, 2011)have
3. This fact can be seen as an absolute ar-
(Suh et al., 2007)
on 6 June 2012)
studies of Wikipedia (Retrieved on 6 June 2012)
been devoted to analyze the manner in which users in-
teract and make use of Wikipedia.
Therefore, our main objective is the classification
and categorization of the most solicited content of
Wikipedia, as we consider that this is directly related
to the use given by users to the Encyclopedia. Unlike
based on surveys conducted for specific populations,
such as (Konieczny, 2007) or (Willinsky, 2007), our
methodological approach is based on the analysis of
a sample of the requests sent to Wikipedia by users.
With this, we consider that the focus of our analysis
is significantly widened: users, Wikipedia editions,
The rest of the paper is organized as follows.
Next section presents the most relevant aspects of the
methodology conducted to develop our analysis. Af-
ter this, some important results are adequately pre-
sented. Finally the section dedicated to present our
conclusions and ideas for further work finishes the pa-
The Wikimedia Foundation has deployed a layer
of special Squid servers to deal with the incom-
ing traffic directed to its several wiki-based projects.
Squids work as reverse proxy servers, and perform
web caching to avoid the operation of both web and
contents can be found in their caches, users receive
their response directly from the Squids. Otherwise,
Squids ask web servers for the solicited resource and,
once obtained, they send it to the user. Independently
of the way in which contents sent to users are ob-
tained, Squid servers write down in a log line data
related to each particular served request.
The analysis presented here is based on a sample
of the log lines stored by Squid servers. All the Squid
servers arranged by the Wikimedia Foundation pack
and send their log lines to a central aggregator host.
In this system, lines are managed by special log pro-
cessors that can write them to a given destination or
pipe them to other processes. In both cases, a sam-
pling factor is configured to determine the number of
lines to be processed. Our sample corresponds to the
whole 2009 and a sampling factor of the 1% was used
to extract it. As a result, we managed to receive one
in every hundred requests composing the traffic to the
several projects maintained by the Wikimedia Foun-
dation. To summarize, more than 14,000 million log
lines have been analyzed as a part of this work.
Receiving log lines from a centralized system is
specially relevant for our analysis as it means that
our sample is made up of lines from all the Squids
deployed by the Wikimedia Foundation as a part of
its Content Delivery Network (CDN), which is com-
posed, at a glance, by two large groups of Squid clus-
ters (located, respectively, in USA and in the Nether-
lands) to whom requests are directed using geograph-
ical DNS balancing policies. This guarantees the het-
erogeneity of our data feed and prevents our results
from localized effects or trends due to sociocultural
particularities that may arise if we only examined log
lines from particular Squids or groups o them.
Once the log lines have been received in our sys-
tems, they become ready to be analyzed by the tool
developed for this purpose: The WikiSquilter project
5. The analysis consists in a characterization per-
formed in a three-step process: parsing, filtering and
storage. Firstly, log lines are parsed to extract rele-
vant informational elements from the users’ requests.
Secondly, these informational elements are filtered to
determine if the corresponding requests fits the di-
rectives of the analysis. Finally, information fields
from requests considered of interest are normalized
and stored in a database for statistical examinations.
Important information concerning users’ requests,
like their date or if they led to a write operation in the
database, can be directly obtained from the log lines
fields. Nevertheless, most of the data needed for our
analysis is embedded in the URL constituting each
request. Therefore, URLs have to be parsed in order
to extract their relevant informational elements:
1. The Wikimedia Foundation project to which the
URL is directed.
2. The corresponding language edition of the
3. When the URL requests an article, its namespace.
4. The title of every requested article.
This information allow us to isolate the requests
directed to the Wikipedia project from all the ones
that compose our sample, as we have just focused on
targeting specific editions. Specifically, we have con-
sidered only the top-ten editions regarding both their
number of articles and visits. These editions are the
German, English, Spanish, French, Italian, Japanese,
Dutch, Polish, Portuguese and Russian ones.
As the articles’ titles are available, we obtained
the ones corresponding to the 65 most visited and
contributed articles in the German, English, Spanish
and French Wikipedias during 6 random months from
2009, and assigned them to an specific set of cate-
gories based in a previous one described in (Spoerry,
5http://sourceforge.net/projects/squilter (Retrieved on 6
2007). At the moment, this is the only step that has to
be performed manually, as there are no categorization
systems that can be fed only with the few words form-
ing a title. The list of categories is presented below:
• Entertainment (ENT) and Current Issues (CUR).
• Politics and War (POL) and Geography (GEO).
• Information and Communication Technologies
• Science (SCI) and Arts and Humanities (ART).
• Sexuality (SEX).
3 Analysis and results
This section presents our most remarkable find-
and contributed articles in the considered Wikipedia
editions. Such results can be considered as represen-
corresponding communities of users. Regarding vis-
its as those requests devoted just to get the informa-
tion contained in the Wikipedia’s articles and that do
not entail any kind of contribution nor any other type
of action, Table 1 presents the categories of contents
most repeatedly demanded in the analyzed editions.
According to this table, we can see how in all the
studied Wikipedias, except the Spanish one, the cate-
gory related to Entertainment topics attracts most of
the visits. However, in the Spanish edition, Scientific
articles are the ones that attract most of the users’ at-
tention. We estimate that this fact could be related
with the high percentage that ICT-related Information
and Communication Technologies contents grab in
the Spanish Wikipedia which is considerably higher
than in the other analyzed editions. As the Entertain-
ment category corresponds to those topics related to
movies, celebrities, video games, music bands, etc.,
results from Table 1 could suggest that Wikipedia is
not considered as a primary source for academic or
scientific information by users in the editions where
this category is the most popular one. Considering
the common range of ages to which these types of
contents are directed to, these results could also al-
low to infer that a large number of Wikipedia vis-
itors in these editions are young people. This fact
could be reinforced by the great percentage achieved
by articles related to sexual topics in some of the
Wikipedia editions with the highest percentages in
Entertainment-related contents. Unfortunately, it is
not possible to go beyond these mere speculations
with the data currently available to us, as Wikimedia
Foundation’s strong policies about individual’s pri-
vacy and confidentiality rights do not allow us to get
Table 1: Categorization of the 65 most visited pages in the
German, English, Spanish and French Wikipedias.
any demographical information about Wikipedia vis-
itors. On the other hand, Table 2 presents the cat-
egorization resulting from classifying the most con-
tributed articles in the same Wikipedias. As it can be
shown, articles related to the Entertainment category
gather most of contributions in the English and Span-
ish Wikipedias. In the English edition, articles related
to current events rank in the second position, whereas
in the Spanish edition the same position is occupied
by articles corresponding to geographical topics. In
the German edition, articles devoted to current events
receive the highest number of contributions whereas
Humanistic articles are in the second position. Fi-
nally, in the French Wikipedia articles related to Hu-
manities are the most contributed while those corre-
sponding to Entertainment are in the second position.
It is common to assume that Entertainment and
Current topics should be the ones most easy to con-
tribute to as they do not require users to deal with aca-
demic subjects nor a previous education, training or
context on a given matter. However, it is really in-
teresting to note how, for example, though scientific
articles are the most popular in the Spanish edition of
Wikipedia, is in the German one when they achieve
the highest percentage of contributions. This means
that users of the Spanish Wikipedia do consume a lot
of scientific information but the contributions, in the
opposite, do not target at all the same subjects.
As we can see, there are significant differences not
and contributions in the analyzed Wikipedia editions,
but also in the topics most visited and edited for each
edition individually considered. This gives us an idea
of the different types of use given to the Encyclopedia
in the different communities of users and the different
patterns of use depending of the considered kind of
request even inside a particular edition.
Category Download full-text
Table 2: Categorization of the 65 most contributed pages in
the German, English, Spanish and French Wikipedias.
4 Conclusions and future work
As our analysis revealed, there are considerable
differences among the types of contents most repeat-
edly visited and contributed in the different editions
of Wikipedia. If we focus on particular editions, we
can, even, found significant differences amongst the
topics that grab users’ attention and those that receive
the most of the contributions. This can be regarded as
different patterns of use characterizing the utilization
made of the Encyclopedia by the different communi-
ties of users. Such an analysis of behavioral features
would benefit from the inclusion of other aspects such
as sociological or sociocultural considerations.
In addition, we could extend our analysis with
the development of an automatic categorization sys-
tem capable of performing a wide-scope classifica-
tion. Nowadays, some of our best efforts are directed
towards this target. In addition, topics involved in
other types of requests, such as history reviews or
searches, could be also categorized to obtain a more
complex profile of requested contents. For sure, some
form of users related information, though hashed or
anonymized, will contribute to define more accurate
Capocci, A., Servedio, V. D. P., Colaiori, F., Buriol, L. S.,
Donato, D., Leonardi, S., and Caldarelli, G. (2006).
Preferential attachment in the growth of social net-
works: the case of wikipedia.
Chesney, T. (2006). An empirical examination of
wikipedia’s credibility. First Monday, 11(11).
Giles, J. (2005). Internet encyclopaedias go head to head.
Kittur, A., Suh, B., Pendleton, B. A., and Chi, E. H.
(2007). He says, she says: conflict and coordination
in wikipedia. In CHI ’07: Proceedings of the SIGCHI
conference on Human factors in computing systems,
pages 453–462, New York, NY, USA. ACM Press.
Konieczny, P. (2007). Wikis and wikipedia as a teaching
tool. International Journal of Instructional Technol-
ogy & Distance Learning, 1.
Korfiatis, Nikolaos, Poulos, Marios, Bokos, and George
(2006). Evaluating authoritative sources using social
networks: an insight from wikipedia. Online Informa-
tion Review, 30(3):252–262.
Kuznetsov, S. (2006). Motivations of contributors to
wikipedia. SIGCAS Comput. Soc., 36(2).
Nov, O. (2007). What motivates wikipedians? Commun.
Ortega, F., Gonzalez-Barahona, J. M., and Robles, G.
(2007). The top ten wikipedias: A quantitative analy-
sis using wikixray. In Proceedings of the 2nd Interna-
tional Conference on Software and Data Technologies
(ICSOFT 2007). INSTICC, Springer-Verlag.
Priedhorsky, R., Chen, J., Shyong, Panciera, K., Terveen,
L., and John (2007). Creating, destroying, and restor-
ing value in wikipedia. MISSING.
Reinoso, A. J. (2011). Temporal and behavioral patterns
in the use of Wikipedia. PhD thesis, Universidad Rey
Juan Carlos. http://gsyc.es/ ajreinoso/phdthesis.
Reinoso, A. J., Ortega, F., Gonzalez-Barahona, J. M., and
Herraiz, I. (2010). A statistical approach to the im-
pact of featured articles in wikipedia. In International
Conference on Knowledge Engineering and Ontology
Development, Valencia, Spain.
Spoerry, A. (2007). What is popular in wikipedia and why?
Suh, B., Chi, E. H., Pendleton, B. A., and Kittur, A.
(2007). Us vs. them: Understanding social dynam-
ics in wikipedia with revert graph visualizations. In
2007 IEEE Symposium on Visual Analytics Science
and Technology, pages 163–170. IEEE.
Suh, B., Convertino, G., Chi, E. H., and Pirolli, P.
(2009). The singularity is not near: slowing growth
of wikipedia. In WikiSym ’09: Proceedings of the 5th
International Symposium on Wikis and Open Collab-
oration, pages 1–10, New York, NY, USA. ACM.
Tony, S. and Riedl, J. (2009). Is wikipedia growing a longer
tail? In GROUP ’09: Proceedings of the ACM 2009
international conference on Supporting group work,
pages 105–114, New York, NY, USA. ACM.
Urdaneta, G., Pierre, G., and van Steen, M. (2007). A de-
centralized wiki enginge for collaborative wikipedia
hosting. In Proceedings of the 3rd International Con-
ference on Web Information Systems and Technolo-
gies, pages 156–163.
Vi´ egas, F. B., Wattenberg, M., Kriss, J., and van Ham,
F. (2007). Talk before you type: Coordination in
wikipedia. In MISSING, pages 78–78.
Willinsky, J. (2007). What open access research can do for
wikipedia. First Monday, 12(3).