Which language edition of Wikipedia has more registered women?

A paper of mine published in 2014 started with this simple (but interesting, I think) question ;)
As you might know, Wikipedia is not available only in English but there are almost 300 Wikipedias written in other languages.
So what we did? We computed the percentage of females and males among registered users on 289 language editions of Wikipedia.

The pdf of “Gender Gap In Wikipedia Editing: A Cross-Language Comparison” is available for you to read.
But I suggest you to try to answer to the following questions beforereading the answers (which are in the paper) so that you might play a bit with your stereotypes and prejudices about culture and women around in the world ;)

1) Which language edition of Wikipedia has the largest percentage of registered users setting their gender as female? What is this percentage? It is more or less than 50%?
And 2) what is the language of the Wikipedia with the smallest percentage of women? How close to 0% might this be …?
3) Try to order the following language editions of Wikipedia from the largest percentage of female registered users to the smallest: Arabic, Bulgarian, Catalan, Chinese, French, German, Hindi, Japanese, Korean, Persian, Swedish, Thai. Where does the largest Wikipedia (the English one) is placed?
4) Moreover, considering that setting the gender on Wikipedia is optional and actually few users do it (see details in the paper). Which percentage of users set their gender on English Wikipedia? What is the Wikipedia in which most users set their gender? What is this percentage?

Note that, as written in the paper, of course languages do not map directly to countries. For example, Spanish Wikipedia is heavily edited from Spain but also Latin America and a similar point can be made from Arabic Wikipedia. India has many official languages Hindi, Bengali, Malayalam, Tamil, Marathi but also English. On the other hand, Italian Wikipedia or Catalan Wikipedia are much more “localized”.
Note also that in the paper we arbitrarily decided to consider only editions with at least 20.000 registered users since we computed percentages on registered users (a Wikipedia with 2 users setting their gender would have had percentages of 0%, 50% or 100% clearly not informative) and this filtering step reduced our sample to 76 Wikipedias with a large number of registered users (at least 20.000).
Note also that data refers to March 16, 2013 but we released the Python script as open source so you can re-run it if you are curious about the current situation. You can get the script on Github.

Ok, now you can go to read the paper “Gender Gap In Wikipedia Editing: A Cross-Language Comparison” to get the answers to the previous questions and hopefully be amazed! Enjoy! ;)

Manypedia presented at Wikisym

I uploaded on slideshare the presentation I gave time ago at Wikisym 2012. It is embedded below. There is a comparison of Points of View across different wikis (such as ecured.cu, the Cuban government official wiki, and Conservapedia, Anarchopedia, veganpedia, …) and a comparisons of the same page across different language Wikipedias thanks to Manypedia (such as “List of controversial issues” in English, Chinese and Catalan Wikipedia, “Human rights in the United States” in English and Chinese, “Osama Bin Laden” in English and Arabic, “Vietnam War” in English and Vietnamese, “Northern Cyprus” in Turkish, Greek and English Wikipedia, “Underwear” in English and Arabic)

Manypedia is online at http://www.manypedia.com.

The paper is at http://www.gnuband.org/papers/manypedia-comparing-language-points-of-view-of-wikipedia-communities/. If you like Manypedia and the paper, please cite it. Thanks!

Which Wikipedia pages are edited mainly by females?

Some time ago we developed Wikitrip, a web tool which shows the world location of editors of a chosen Wikipedia page and also the gender of editors, i.e. how many edits were made by males and females. We released Wikitrip as open source on github and we also deployed 3 live APIs: api.php (various stats about a specific Wikipedia page), api_gender.php (returns timestamp and gender for any edit to a specific Wikipedia page by a registered user that specified his/her gender), api_geojson.php (returns a GeoJSON for anonymous edits on a specific Wikipedia page). If you want to use the APIs in your mashups, we’ll be delighted, more details about the APIs can be found at the end of this blog post.

In fact, today I discovered via Gizmodo that Santiago Ortiz has used our Wikipedia Gender API for creating a fantastic visualization of Wikipedia pages based on how many female and male contributors each of the articles has.
Using the cool visualization you can for example “discover” that currently, out of more than 4 million pages in the English Wikipedia, JUST ONE article is edited more by females than males!!! That article, with 7 male editors and 9 male editors, is Cloth menstrual pad.
Note that the API we released is based on data from Wikipedia and that only users who specified their gender in Wikipedia are counted in (these users are a minority, around 10%). Note also that in our Wikitrip visualization we show a plot with number of edits from gendered users while Santiago show the number of different users. For example in Wikitrip you see that the page Cloth menstrual pad has 62 edits from females and 15 edits from males but the different users who edited the page are 7 male and 9 female.

Wikipedia gender divide visualized

Now I give you some more info about the APIs released with WikiTrip so that you can use them as well in your mashups if you wish so.

  1. api.php

    Get various stats about a page


    • article: page title
    • lang: desidered language (default: en)
    • family: wiki family (default: wikipedia)
    • year_count: show edit count per month (default: false)
    • editors: show unique editors for the page (default: false)
    • max_editors: maximum number of editors displayed (only if “editors” option is set)
    • anons: show anonymous unique editors (default: false)
    • top_ten: show top 10% of editors (default: false)
    • top_fifty: top 50 editors (default: false)


  2. api_gender.php

    Get timestamp and gender for any edit by a registered user that specified his gender on a specific page (might be quite slow)


    • article: page title
    • lang: desidered language (default: en)
    • family: wiki family (default: wikipedia)


  3. api_geojson.php

    Get a GeoJSON for anonymous edits on a specific page


    • article: page title
    • lang: desidered language (default: en)
    • family: wiki family (default: wikipedia)


P.s.: the coder of all Wikitrip awesomeness is the amazing Federico “fox” Scrinzi!

Models of economic production, Encarta vs Wikipedia, and sober economists

An excerpt from Dan Pink’s TED talk “The surprising science of motivation

The mid-1990s, Microsoft started an encyclopedia called Encarta. They had deployed all the right incentives, all the right incentives. They paid professionals to write and edit thousands of articles. Well-compensated managers oversaw the whole thing to make sure it came in on budget and on time. A few years later another encyclopedia got started. Different model, right? Do it for fun. No one gets paid a cent, or a Euro or a Yen. Do it because you like to do it.

Now if you had, just 10 years ago, if you had gone to an economist, anywhere, and said, “Hey, I’ve got these two different models for creating an encyclopedia. If they went head to head, who would win?” 10 years ago you could not have found a single sober economist anywhere on planet Earth who would have predicted the Wikipedia model.

MIT Press book “The Reputation Society” (containing a chapter by me) is out!

The MIT Press I contributed to with a chapter is out! It is titled “The Reputation Society: how online opinions are reshaping the offline world” and edited by Hassan Masum and Mark Tovey.
It is available on MIT press and on Amazon.
The chapter I wrote is titled Trust It Forward: Tyranny of the Majority or Echo Chambers? and on it I ramble about objectivity/subjectivity, minorities/majorities, etc.

If reputation systems weight all perspectives similarly, they may devolve into simple majority rule. But if they give each user reputation scores that take only other similar users’ opinions into account, they run the risk of becoming “echo chambers” in which like-minded people reinforce each others’ views without being open to outside perspectives. Massa discusses design choices and trust metrics that may help balance these two extremes and the broader implication for our future societies.

the reputation society book cover The book received endorsements by people I really admire.
“As our societies expand from local villages to global networks, our ways of assessing and sharing reputation—the foundation of trust and community—must also evolve, but how? The thoughtful and thought-provoking essays in The Reputation Society bring a wide range of perspectives to this question, including the design of technological solutions, applications in philanthropy, science and governance, and warnings about the loss of privacy and autonomy. It is a fascinating collection of readings not only for scholars, but for anyone interested in the dynamics of the reviews and recommendations that shape our decisions—or in the future of how we will judge and be judged.”
Judith Donath, Fellow, Berkman Center for Internet and Society, Harvard University

“Today is tomorrow’s yesterday. These provocative essays, by some of the leading thinkers in the domain of reputation systems, illuminate how reputations regulate actions across time and social distance and point to the opportunities and obstacles that reputation systems present for commerce and democracy.”
Paul Resnick, Professor, University of Michigan School of Information

“The Reputation Society enriches the discussion of reputation by bringing together technologists, philosophers, legal scholars, and industry leaders to sort through the promise and perils we face today. It covers the practical, for those interested in the nuts and bolts of the challenges we face today, and the theoretical, for those looking to engage in broader discussions of the ethical and moral concerns. In short, a terrific and enlightening read!”
Danielle Keats Citron, Professor of Law, University of Maryland School of Law

The list of my co-authors is also very delightful.
Trust, reputation systems, and the immune system of democracy / Craig Newmark
Building the reputation society / Hassan Masum, Mark Tovey, & Yi-Cheng Zhang
Designing reputation systems for the social web / Chrysanthos Dellarocas
Web reputation systems and the real world / Randy Farmer
An inquiry into effective reputation and rating systems / John Henry Clippinger
The biology of reputation / John Whitfield
Regulating reputation / Eric Goldman
Less regulation, more reputation / Lior Strahilevitz
The role of reputation systems in managing online communities / Cliff Lampe
Attention philanthropy : giving reputation a boost / Alex Steffen
Making use of reputation systems in philanthropy / Marc Maxson & Mari Kuraishi
The measurement and mismeasurement of science / Michael Nielsen
Usage-based reputation metrics in science / Victor Henning, Jason Hoyt, and Jan Reichelt
Open access and academic reputation / John Willinsky
Reputation-based governance and making states “legible” to their citizens / Lucio Picci
Trust it forward : tyranny of the majority or echo chambers? / Paolo Massa
Rating in large-scale argumentation systems / Luca Iandoli, Josh Introne, & Mark Klein
Privacy, context, and oversharing : reputational challenges in a Web 2.0 world / Michael Zimmer & Anthony Hoffman
The future of reputation networks / Jamais Cascio
“I hope you know this is going on your permanent record” / Madeline Ashby & Cory Doctorow.

The cover of the book reads as follows.

In making decisions, we often seek advice. Online, we check Amazon recommendations, eBay vendors’ histories, TripAdvisor ratings, and even our elected representatives’ voting records. These online reputation systems serve as filters for information overload. In this book, experts discuss the benefits and risks of such online tools.

The contributors offer expert perspectives that range from philanthropy and open access to science and law, addressing reputation systems in theory and practice. Properly designed reputation systems, they argue, have the potential to create a “reputation society,” reshaping society for the better by promoting accountability through the mediated judgments of billions of people. Effective design can also steer systems away from the pitfalls of online opinion sharing by motivating truth-telling, protecting personal privacy, and discouraging digital vigilantism.

Article about Manypedia on Italian newspaper Corriere

I’ve been interviewed by the Italian newspaper Corriere della Sera about Manypedia (and Wikitrip). If you know Italian, you can read the resulting article titled “Every Wikipedia represents its own culture: even the concept of controversiality is controversial” at corriere.it. The journalist liked to stress the fact both Manypedia and WikiTrip are open source, which is a good thing I think.
Manypedia on corriere.it

My presentation at Wikisym: studying (current) history by analyzing Wikipedia

Wikisym was a great conference! Below you can find my presentation about the paper on Collective memory building in Wikipedia. During the presentation, I provided evidence and possible research lines in order to argue how it is becoming possible to study history (of current events) by analyzing what it is written about these events by thousands of editors on Wikipedia.

WikiTrip: animated visualization over time of gender and geo-location of Wikipedians who edited a page

WikiTrip allows to have a trip in the process of creation of any Wikipedia page from any language edition of Wikipedia. WikiTrip is an interactive web tool empowering its users by providing an insightful visualization of two kinds of information about the Wikipedians who edited the selected page: their location in the world and their gender.

If you want to investigate, for example, where in the world are Wikipedians who edited the page “Peace”, WikiTrip is the right tool. And you can check also the origin of edits for the equivalent page in the Arabic Wikipedia or “Amani” in the Swahili Wikipedia. Moreover, if you have ever wondered if a specific page was edited more by male or female Wikipedians, WikiTrip allows to explore this information as well. How many edits are performed by males and females respectively on Wikipedia on average? What is the page most edited by females? On Wikirip you can explore your own ideas about these questions and more.
Visualization of both information is available over time so that you can appreciate the evolution of the page over years, from its creation up to the present.
More information about WikiTrip at our whitepaper but the best way to enjoy WikiTrip is at http://sonetlab.fbk.eu/wikitrip/.
We would love to hear the Wikipedia pages you found more interesting as they are visualized by WikiTrip and of course we wait for your feedback!

Wikisym and collective memory building of current events on Wikipedia

In few hours I’ll start the long journey towards Mountain View, California, for the Wikisym conference where I’m going to speak about WikiRevolutions presenting the paper “Collective memory building in Wikipedia: The case of North African uprisings“.
In the paper, we highlight the intense edit activity by Wikipedians on articles related to protests and uprisings in North African countries such as Tunisia, Egypt, Libya, Syria, Yemen focusing mainly on the Egyptian revolution.
We cast the phenomenon as a process of collective memory building in which thousands of Wikipedia editors were involved as the traumatic events unfolded.
We explore and suggest possible directions for quantitative research on collective memory formation of traumatic and controversial events in Wikipedia.
I’m in a Wikisym session titled “Wikipedia as a Global Phenomenon” in which I will have the pleasure to speak after Brian Keegan that is addressing the same topic of how on Wikipedia it is possible to analyze how editors cover recent events in real time; the paper is “Dynamics, Practices, and Structures in Wikipedia’s Coverage of the T?hoku Catastrophes” (joint work with D. Gergle, N. Contractor).
Probably at Wikisym there will be also people from Ushahidi, such as Heather Ford, which recently announced WikiSweeper, a joint project with the Wikimedia Foundation to track breaking news trends on Wikipedia so I think we will have wonderful exchanges of points of views and possibly future collaborations.
I’m really looking forward for what looks like a fabolous conference!

Credit: The image on top is Creative Commons BY-SA.

P.S.: At Wikisym I’m also going to demo our web tool on comparing points of view of different language communities of Wikipedia, i.e. Manypedia.

FBK researcher awarded by Google with 50,000 dollars and … emotion ;)

One of my colleague at Fondazione Bruno Kessler (FBK) in Trento, Carlo Strapparava, was awarded with $50,000 by Google as an incentive to continue his research, especially with the participation of young researchers. Carlo proposed algorithms for distinguishing some of the nuances and emotions expressed in written language.

Dealing with the emotional, persuasive, or other aspects of creative language content in the texts – Strapparava explains- is commonly considered to be off limits for any computational ability. Actually, these features are a key part of communicating, and it is important that research in the field of natural language processing deal with it. The usefulness of automatic recognition of these aspects is nowadays even greater, given the enormous daily production of texts on the web. Through these technologies, it will also be possible to predict the emotional or persuasive content of a text.

Congrats Carlo!