The most popular posts
in the last 7 days
- WoW loses another 100,000 subscribers 156 comments
- The Daily Grind: What dead game would you play in a second if it revived? 99 comments
- The Daily Grind: What's the highest sub fee you'd pay? 86 comments
- BioWare kicks off Star Wars: The Old Republic Q&A Fridays 78 comments
- Earthrise shutting down today 69 comments
Massively Speaking Podcast
Massively Speaking Episode 185: Bree-to-play
Latest episode: Tuesday, February 7th, 2012



Reader Comments (2)
Posted: Feb 16th 2009 4:11PM (Unverified) said
1. predict who was likely to cancel their subscription
2. build a predictive model to identify gold seller accounts
3. build another model to identify gold buyers
4. identify community influencers through their interactions with others
5. cluster customers into similar behaving groups for the purpose of understanding behavioural trends (think Bartle test using real rather than claimed behaviour)
6. use econometrics to predict network usage patterns
7. maybe some influence and sentiment analysis of the free text
All of these analyses are pretty simple to conduct apart from the last one. The data would need some flattening and interactions categorised - which would be 80% of the effort.
Posted: Feb 16th 2009 8:48PM Graill440 said
As for the "Analyst" part, what does anyone hope to gain by reading player chat logs and the like, by showing how ignorant and banal people can be when they are behind a monitor?
There are plenty of studies showing subscription trends and server population migrations, why do something like this again, to waste my taxpayer money? It certainly isnt coming from a rich private grant.
The waste and ignorance of the people conducting this study simply reflects a want for some type of job security to get paid so they dont hit the street and souplines, wasteful and uneeded. Spend the money and time on something worthwhile like improving your community.