Screenshot / Code Snippet
StackTagz analyses your participation in StackExchange sites by looking at which tags you are active in. A question, an answer or a comment is counted as an "interest" point in a tag. The results are rolled into weeks as this period was found to give the best overview of one's activity.
It uses a local database to cache API results fairly aggressively. Because it is assumed that once a question is out there, it's tags don't really change, it is possible to cache full timeseries and individual questions. The goal is to have the app not making a repeated call to the API ever.
Version 1.0. Future revisions will include better feedback about the progress of the graph request and turning individual timeseries on and off.
Public website, free to use.
Contact me at email@example.com if you have any queries or suggestions. I am planing to set up a blog and a twitter account for this site, so stay tuned.
At this stage the code is closed source as it is a public website that requests makes A LOT of API calls. Once I am confident that it's fairly secure, I am planning to open source it.
Of course, I will provide source code upon request if it so happens the app is considered for the API Contest :)
Update (2010-07-06) User search is now performed using the api server directly, instead ofusing the stacktagz server, which in turned queried the API. Suggested users should come up much quicker now.
Update (2010-07-08) Fixed tooltips not being shown or shown incorrectly for some points on the graph. Added site links for user graphs.
- Comments are now processed.
- Fixed a bug where an answer to own question will count twice towards the answer.
- Changed Rep display for 10,000-100,000 range when the decimal part is 0 - i.e. 12k instead of 12.0k.
- Changed Rep to show 'k' instead of 'K'.
Just as I am writing this and getting graphs for above users, the results are fascinating. People's interests are absolutely fascinating!!! Compare cletus (who is into everything) and Alex Martelli (die hard pythonista). Or that shog9 is really Jeff Atwood's alter ego (look at the pattern in the first year). Or that Jon Skeet gradually lost interest in Meta (tut tut Jon :D ).
I am thrilled to see this kind of data side by side.