Interview: WowLens and a National Science Foundation grant help researchers mine player data

Addons have been used in World of Warcraft for a variety of in-game purposes, such as data monitoring or making players' lives easier. What happens, however, when smart people decide to use the power of WoW's interface for a scientific and statistical purpose?

Researchers at the University of Minnesota, Carnegie Mellon University and the University of Pittsburgh, under a project by the National Science Foundation, are using an addon called WowLens to gather peer review data. The project aims to use the wealth of people, resources and data floating around Azeroth in compiling data for research projects. It personally reminds me of Folding@home, but with statistics instead of computer processor cycles for medical calculations.

I got to have some questions answered about the project by Professor John Riedl of the University of Minnesota Computer Science department and Daniel Moy, one of his colleagues and students. If you would like to get involved with the project, simply head over to the WowLens site and download the client. You run the program before opening World of Warcraft, answer a few survey questions, and at the completion of a group, you answer a set of questions pertaining to your experience with the players in that group. Let's let the researchers themselves tell you about their project.

AddOn Spotlight: Can you give wow.com readers a brief synopsis of the project you are working on? What are the aims of the study?

Professor John Riedl: My group is GroupLens Research at the University of Minnesota. You can read our Wikipedia page here. We wrote much of the page, but it's the most complete telling of the story.

In the past seven or so years, we've been working with our colleagues from the social sciences at the University of Michigan, Carnegie Mellon University and the University of Pittsburgh, to apply insights from how groups of people work together to improve online collaborations. For the past three years, we've been using World of Warcraft as one of our study platforms. We're particularly interested in developing new intelligent user interface technologies that can make teams in WoW work together more effectively.

The current plug-in is one of those pieces of research. We're exploring how what is known about group formation from the social psychology literature can be applied to help make WoW groups happier and more effective at their goals.

Daniel Moy: The overarching research we're working on is a National Science Foundation grant between three universities (Carnegie Mellon University, University of Minnesota, University of Pittsburgh), looking at online groups. This covers groups in a variety of different settings, including both Wikipedia and World of Warcraft. Specifically for World of Warcraft, we're looking at figuring out how to understand and improve group satisfaction, cohesion and coordination. Finally, WowLens itself is looking at whether or not we can use an in-game rating system (of peer ratings, not arena ratings!) to help players form better groups and in general have more successful, happy game experiences.

AS: What interested you guys about Wow? Was it the size of the player base? Are members on the team players?

Moy: Yes, the large player base size is a definite attraction. If millions of people use a system, chances are there could be interesting research questions to answer. There's also quite a bit of existing academic research into various aspects of World of Warcraft. As for the second part of the question, most of the group has played WoW at one point in time, including one of the professors. Only two of our current group members are ACTIVE players, though a couple more keep up with the important news, trends, etc.

AS: What were the contributing factors to choosing WoW as the testing ground?

Professor Riedl: We're particularly intrigued by WoW because of the team aspects of the game. We like the way small "work groups" (parties) and larger "organizations" (guilds) interact over time. One of the challenges for WoW is how these small groups have to form over a short period of time, and then fade away, replaced by other small groups. These sorts of dynamic reformings are unusual in physical world groups, but seem much more common in online communities. What tools will best support this new community behavior? How can it be made more fun and effective?

Several of us are players. Daniel has a long history with the game. Tony (the lead grad student here at Minnesota), Bo (the lead grad student at Carnegie Mellon University) and I have also played, though not to the expert level. Kristina (the lead grad student from Pitt) and Vanessa (a M.S. student from Minnesota) have played to end game. On a personal note, my son Kevin has played to the expert level, and I've enjoyed talking about the game with him.

We are very impressed with the game as an enormous-scale online community. The level of interaction among players is fun and interesting, and the scope of the world is breath-taking.

Moy: MMOs are useful in their own right for a many reasons. Some of these include complex or obscure task solving, spontaneous group forming and the insane amount of time users tend to play. World of Warcraft specifically does have a alluring amount of research data, with orders of magnitude more players than other MMOs.

AS: Who developed the addon for you guys?

Moy: I did, in its entirety. I played as Pharaunmizz on Greymane and had a bit of an addon affection complex already going when we started. When we were looking for ways to get research done, we tossed around the idea of using an addon. I had to learn Lua and the WoW API to actually implement it all, which was a long and painful process. Big, big thanks to wowwiki, wowprogramming and #wowuidev.

Professor Riedl: We're impressed with the Lua environment for extending the WoW interface and would love to build further add-ons to make players' WoW interactions more positive. We'd love to hear any ideas you have for addons that guilds would appreciate!

AS: What can readers do to help contribute to the work you do?

Professor Riedl: The most important is to use the addon! We're very excited to get user feedback and learn how to make the addon more valuable.

We'd particularly love the opportunity to work closely with some of the large guilds to build tools with and for them that would help their interactions. Please encourage your readers to contact us at wowlens@cs.umn.edu. The great thing is that we have research support to build cool tools for WoW players. Let us know what the community would value, and we'll figure out if there are interesting research questions to explore!

Moy: The first is to use our addon, WowLens, and religiously submit data. We've built a clean little addon that runs in-game like any other addon, and a data uploader that will upload data to our servers, download an updated data set for you, sync everything up and automatically start World of Warcraft. This uploader is meant to be run instead of the usual WoW shortcut to start the game; it should handle starting the game itself cleanly for the user.

AS: What other uses do you think addons can haev in research besides data collection? Any other impacts on the research world?

Moy: A good question! Passive data collection, while useful, isn't as fun and can't answer a lot of the interesting research questions. The potential for using addons in active research experiments is very, very appealing. Also, simply having a user base with an addon linked to a research community, perhaps on an outside website, could be invaluable, especially if it gains any popularity with the player base. Again, if anyone has any thoughts about future directions, we're always willing to listen and explore new ideas.

Professor Riedl: Over the long-term we'd like to create tools that make peoples' social lives richer and more fulfilling. We are enthusiastic about WoW as an online community with a powerful social component, and would love to explore how research can make it even more satisfying.

Once again, thank you to Professor Riedl and Daniel Moy for taking the time to answer my questions. If you'd like to help out with the Wowlens project, check out their website and provide some feedback!