SourceForge Stats Demystified
Hi! In my job, I get to to travel the world and talk to a lot of SourceForge.net users. One question always keeps popping up: what, exactly, does the “activity percentile” actually mean? Or, more to the point: why is my activity percentile lower than another project that has usage patterns that seem to exactly match mine? Read on to learn more about the tool that can help you answer these questions!
People are often surprised to learn that our formula for calculating a project’s activity percentile is public, and can be found in our Site Docs. Essentially, it’s a combination of three different classes of activity: traffic, development, and communication.
The traffic class contains the number of hits to the project’s pages on SourceForge.net, the number of logo hits, and the number of downloads. The development class is comprised of the number of CVS or SVN commits, the number of days since the last file release, and the number of days since the most recent project admin login. Communication currently contains the number of new Tracker submissions and the number of forum posts.
These three sets of numbers get compared to those of other projects, and eventually become the relative activity percentile.
Now comes the really good part. You can see the formula in action on a project’s Rank History page, which contains the values of each of these numbers for a particular day in the recent past. It’s a bit hidden, for the next few weeks or so, but you can still get to it through the Software Map - just click on the number that shows up in the Rank column for your project. The screenshot to the right shows where that link is. To see what an example Rank History page looks like, check out the one for Gallery.
You can also click on one of the date fields shown on the Rank History page to see the formula itself with the numbers for that day’s activity plugged in! What will they think of next?
Enjoy your new stats obsession,
Ross
Reader Comments
open formula realy comes as a surpise 
gr8!
just a few questions on this formula:
1) What do you think, how much the typical project user can rely on this ranking? Is it accurate enough to make a choice of software based on this?
2) As I understand Traffic, Development and Communication have equal weights in your formula. Would it makes sence to add unequal weights to different parts of the formula. On my opinion, communication could give much more to the score than pure traffic of website. Or even better, would be to give user a possibility to assign weights he or she thinks are most suitable. What do you think?
3) Do you plan to introduce other indicators in addition to activity? Would be great to rank projects by their popularity, quality, …
good luck,
Mark
Bio: http://www.sourcekibitzer.org/Bio.ext?sp=l8
Hey Mark!
1) This ranking is intended to be a measurement of the project’s overall activity. A high level of consumption suggests that a project has a high level of maturity, but it’s not a guarantee of quality. I think it’s one of many factors users should consider when choosing software to meet their needs.
2) I think it depends on what you’re looking for! I think that weighing them all equally now is the most fair way to do it, and provides the most level playing field for our projects. However, the idea of allowing each user to specify what’s important to them is very interesting to me (even though it would probably be very difficult to implement.)
3) We don’t have any current plans to do anything like that, but I would appreciate any ideas people have!
Thanks, Mark. Nice bio! That SourceKibitzer thing looks neat. 
Ross
I think it would be fair to include mailing list posts in the communication class of activity.
–
cheers
nek
npapadop: For what it’s worth, including mailing list posts in the stats system is on the slate of work to be put into the stats/ranking system in the near future, though we have some other tweaking which needs to take place first.
phuff, reluctant captain o’ stats.
Hmmm, strange. I do login every day as projectadmin to have a look at the bugtracker in our project. It does login automaticly (we do remember you, is SF saying) but still the login stats say it is 17 day’s for the last login!!
What kind of login is measured in the stats? Or is there a little bug inhere?
Dear Ross,
I am sorry to bug you, but we recently installed the sourceforge forum system on our server (MS SQL 2005) and were excited how well it worked.
However, after the most recent MS update, and using the compiler Visual Webdeveloper Express, we are suddenly receiving the error message:
“An attempt was made to load a program with an incorrect format. (Exception
from HRESULT: 0×8007000B)”
Can you help us?
Many thanks,
Robert “Bobby”
Hey Bobby - not sure what forum system you’re talking about. SourceForge.net is a hosting repository with almost 155,000 separate software projects. You should probably locate the project for the forum software you’re using and ask the maintainers. If you need assistance doing so, feel free to drop me a message by going to http://sourceforge.net/sendmessage.php?touser=38148.
Thx,
Ross
And I’m __underlined__!
So, what about projects that don’t use those services, forums and svn/cvs? Do they get a penalty for not using them?
Both comments and pings are currently closed.
Search Community
November 2009
All events for November
Most Active Forum Posters
- silverfang (553)
- hinojosa (355)
- moorman (325)
- creek23 (153)
- trilarion (124)
- leeschlesinger (111)
- caglow (108)
- rturk (79)
- javajox (71)
- bricelambson (68)
Thats really a great idea but dont you think that it would be too tough for an ordinary developer.
Regards,
Komail Noori
Web Designing and SEO Expert