Mobility in an online world
From Santa Fe Institute Events Wiki
|CSSS Santa Fe 2010|
Massive multiplayer online games (MMOG) provide a fascinating new way of observing hundreds of thousands of simultaneously interacting individuals engaged in virtual socio-economic activities. We have compiled a data set consisting of practically all actions of all players over a period of three years from a MMOG played by over 300,000 people. The universe of this online world is a network on which players move in order to interact with other players. This interaction may consist of trade or armed conflict, players can also mark each other as friend or enemy. We focus on the mobility of players on the network over a time-period of 500 days. We take mobility measurements (e.g. entropy) of players and compare them with measures of simulated random walkers. Player mobility is highly different from mobility of random walkers, even when these share similar hop distributions. Further, we measure mobility during times of in-game war and non-war, and uncover different exponents in hop distributions: during war mobility increases.
Massive multiplayer online games (MMOG) provide a fascinating new way of observing hundreds of thousands of simultaneously socially interacting individuals engaged in virtual economic activities. We have compiled a (fully anonymized!) data set consisting of practically all actions of all players over a period of three years from the MMOG Pardus (www.pardus.at) played by over 300,000 people. Here we focus on the movement of players in the game's virtual universe, to shed light on mobility patterns within human societies.
The universe of this online world can be represented as a network on which players move to get involved and to interact with other players. It consists of 400 sectors, where each sector is a rectangular grid made up of 15x15 fields on average. A field is the atomic unit of space on which players move. Each sector has a unique sector-id, each field has a unique field-id, with a corresponding a sector-id. To players, sectors are displayed as nodes on a two-dimensional cartesian coordinate system, the universe map. Nearby sectors -- in the sense of having small distance in the universe map -- may be interconnected by wormhole-links, allowing plyers to hop from sector to sector. While wormhole links always link together exactly two wormholes (i.e. two sectors), so-called X-holes connect to all other X-holes in the universe. This results in a "folding" of space, with X-holes bringing closer far apart sectors. Players have a movement cost for traveling fields and for jumping through wormholes and X-holes, limiting their possible numbers of daily traveled fields and sectors.
- Extract data of universe
- Bring topology data in network format (adjacency list / sparse matrix)
- Visualize network, e.g. in PAJEK, G-PHI
- Extract mobility data of players
- Bring them in reasonable format (time-series). E.g. Time-series of daily hops.
- Analyze data
- Visualize time-evolving data (over the whole time-span: 1. display locations of 5 typical player, 2. display locations of 100 players in war)
- Formulate research questions
- Find literature on mobility, etc.
- Write paper (tex)
Data -> Calculations/Measures
- Time-series of distances travelled (over a time-span) -> Spatial entropy, COV matrix
- Number of unique places visited (over a time-span)
- Time-series of changed friends/enemies -> Social entropy
- How is spatial/social entropy distributed? Do we see modes? If so, why?
- Does mobility change in times of war?
- Do locations and friends/enemies correlate? If so, is there causality?
- Thursday, June 10, 5 PM, lecture room
- Sunday, June 13, 8:30 PM, cafeteria
- Sunday, June 20, whole day, cafeteria
- Tuesday, June 22, 5 PM, cafeteria
- Thursday, June 24, whole day, cafeteria