One of the things I do occasionally is collect some of my own statistics on Drupal.org. An interesting one is to look at the size, expressed in nodes, and the growth rate of Drupal.org, expressed as nodes per hour.
Just like rings on a tree trunk, you can see Drupal.org evolve.
Now while the nodes graph (red, left axis) looks exponential at first, it turns out that its growth (blue, right axis) is actually quite stair-stepped. In fact, a remarkable trend can be observed: with every major release of Drupal, Drupal.org's growth rate doubles nearly instantly.
This has been true mostly since Drupal 4.5 and happened recently again with Drupal 5.0. The only exception was a relatively unexpected, but sustained linear growth burst throughout the fall of 2005 and the spring of 2006. There were many high profile projects at that time as well as a lot of anxious waiting for 4.7 (with a long train of betas and release candidates), so it is possible our unnaturally long dev cycle for 4.7 smeared out the growth to a more regular line.
Note that the first year of Drupal is missing, as the statistics were too small to show up.
What's also interesting is that you can clearly see the times when project.module has changed. The first tiny bump is the conversion of projects to nodes, while the other two spikes are the recent conversion of project releases to nodes. For some reason, a huge amount of existing releases got a creation timestamp a while ago, possibly due to some change in project.module housekeeping at that time.
Other noticeable events include the holidays (when activity slows down a lot) as well as the server crash of July 2005, and the subsequent move to OSUOSL.
I've attached the raw Excel Spreadsheet for anyone interested. See if you can match up any other events to the graph.
| Attachment | Size |
|---|---|
| drupal.org.stats.xls | 688.5 KB |


Bets are on
Here's a fun bet: what will the highest node id be on Drupal.org on January 1st, 2008? You have all the data available to do your own predictions, you could look at the graph, or just plain guess. We'll revisit this thread in a year, and whoever gets closest wins a prize :P (let's say some Belgian beer or chocolate, retrievable at the next DrupalCon after that?).
My guess: 307 000 nodes.
Very interesting
Superb chart design and analytical methodology, thanks Steven. The backdating of project release nodes is particularly smart. The composition of drupal.org nodes is another issue: issues vs. forum posts vs. release nodes, etc.
As the module base grows and becomes less dependent on core functionality (E-commerce v3 for example) it may be increasingly fruitful to analyze contrib commit activity. I took a quick stab last week at examining commits over time, but failed to figure out the cvs commands to extract such data. The cvs mailing list archive is another possible source.
Another issue your chart brings up is the endogeneity of the Drupal release timeline, especially code freeze vice ve actual release. To the extent that it is controllable the release schedule can theoretically be manipulated to achieve higher or lower levels of growth.
Finally, all of the smaller local maxima that I checked corresponded to important Drupal events: beta/RC release dates or code freeze dates. So, that is as one would expect.
about the growth.
Althought i´m happy with the drupal growth i think that a slow buf good growth is better than a fast and furious growth.
I prefer the project were until a think isn´t finished can no begin a new one. This make it to growth up more slowly but at the same time better.
Spanish people are trying to get a good drupal group of people, but since now there ara nothing interesting.
Oskar
Great chart, Steven.
Great chart, Steven.
Drupal
Wow, I knew it was growing but thats awesome. It adds a level of support in my mind that things are going to continue and this is a great thing to be invloved with.
Amazing chart
I wonder how long Drupal.org can sustain such exponential growth? Great care needs to be taken to prevent Drupal's rapid expansion from negatively impacting the quality and usability of the site.
On a similar note, I feel that Drupal's documentation is suffering due to the rapid growth. Things are changing so fast, it's not unusual to find handbook entries that are completely obsolete, usually with a string of comments at the bottom saying "this doesn't work anymore" or "it's broken in 5.x". And 6.0 is already on it's way..
I think Drupal.org needs to take some lessons from Wikipedia and allow free editing of the handbooks. 100,000 users and only 78, less than 0.1%, have access to change anything.
Furthermore, handbook entries should be dated and there should be a way to find out who wrote them.
Drupal Explosion
This chart looks like real explosion. I am looking forward to the new data in January 2008. Let me predict a 240000 nodes and 35 nodes per hour.
How did it go?
So, how many nodes were there January 1, 2008?
Drupal.org node count
Without access to the DB, I can't tell exactly how many public/non-deleted nodes are there, but the first node of 2008 (in the anonymous/default timezone) is http://drupal.org/node/205243
Post new comment