Piwik hits 1.0 - Open source web analytics
piwik.orgI've used piwik, it really powers the way you can do in-house analytics. Real time, javascript based tracking - no need to wait for data to accumulate. Plugin model makes it highly flexible and extensible. Module Design is very good.
I wanted something to analyze existing log data and could quickly write a small extension to do that. Creating new views or data points were straight forward too. Could easily integrating a simple external package like statviz.
Wasn't very fast processing data in bulk - 20 minutes to process 200K visits. But this was like a year ago. Ofcourse php, Zend db has its limitations with data processing and if the fact table and aggregate creation is rewritten in perl, C++ or something and done at predefined intervals without doing a running average it can handle fairly large amounts of data I suppose.
I've been using this for a few weeks and it is really solid. It feels nice to have my analytics data on my own server instead of the GOOG's.
Now they only have access to almost all of my data...
Were you able to import any Google Analytics data into Piwik? Does GA offer a way to get your data out, other than exporting specific reports?
No, I just jumped in cold turkey, but Googling around produced this which you might try:
http://jaymz.eu/2010/02/importing-existing-visitor-stats-fro...
Piwik is nice; another open source tool worth looking at is Open Web Analytics, http://www.openwebanalytics.com/. Both have their advantages and are both under active development.
The real power of GA is watching it scale to handle 100mm+ monthly pageviews. Can Piwik do this?
I've been watching Piwik for a bit (they do analytics software, we do analytics software) and I believe they set aside their scaling work until after version 1.0.
I would love to see an independent open-source analytics company with a services / enterprise business model, along the lines of OpenX for ad serving and the pre-acquisition MySQL. Not so sure the economics justify it, though.
Actually, I've been thinking about doing this for a while -- this thread convinced me to put up my MVP to see what happens.
While I'm waiting for DNS to propogate, here's a sneak peek: http://img.skitch.com/20100831-kf41k4mpc4gd612jyhdcbtyxgh.jp...
edit: it's now available at http://dashboard.io -- would love any feedback I can get. :)
No (speaking from experience - we've had no end of issues with piwik and larger websites)
Could you quantify 'larger'? Just wondering at what point this starts to have problems.
Don't know if this relates directly to the question about scalability, but the main site is not loading for me, nor is the demo site.
It looks like the tracking script hosted on demo.piwik.org is slowing the main site down.
Dealbreaker for me: Piwik cannot do page-specific stats.
Everything else was in Piwik's favor, but I'm going to have to use GA because of this one requirement.
Sounds like there could be a trade-off between using resources on your own server to keep track of these analytics versus just having Google do it. But then maybe it is insignificant.
Anyone has some stats on how much hit on the server it is doing?
Not me, but the privacy trade-off is pretty good. :)
Been using it myself since... around 0.2 or so, very satisfied.
The upgrades were painless too, which was a nice surprise - I expected a few hickups during the earlier releases, but found none.
I've been using it for about 8 months now and I have been nothing but impressed with these guys. It has proven very stable even from their earlier releases and very snappy.
Just wondering: does anyone know of a mere log analysis tool, that would just look at Apache's logs? What data would I miss, compared to Javascript or PHP based tracking?
I'm asking because I'm looking for a non-invasive method, which I could use on old logs if possible.
I used slimstat (w/ PHP) a while back, and it was pretty decent. I wonder how they compare...
An amazing product that I've been following for the last year has reached the 1.0 milestone.
I don't use GA so can't comment on feature parity. Importantly (for me) it has a powerful API, plugin architecture, easy integration, dead easy multi-site configuration.
I can't speak highly enough of this product.