Google Analytics Alternatives?

What’s the best open source alternative to Google analytics and why?

Now that Google has decided it’s against net neutrality, I’m just not willing to use them anymore on any project.

What do you collect for your log files? Do you grab location data? That’s something I’d like to get, but I don’t know of a service that translates IPs into location (at least I assume that’s how its done). Or screen resolution? That would be useful to me, but I don’t even know where to get that.

I’ve looked at a number of open source packages, but I’m more interested in people’s experience now.

Hi Anthony

W3Counter looks to be well featured. I am always concerned about services that appear to be free.

Do you have personal experience of using this? If so just how much better is it the Google Analytics, if indeed it is.

Colin

It’s not a patch on Google to be fair, but it is made my Sitepoint member DanGrossman who seems more than reputable.

Maybe I’m a little naive, but I only put analytics on my public pages and can’t really see what damage can be done by the collection of this user data.

As Tim states though, parsing your log files is a viable option.

Have you seen W3Counter ?

Awstats tends to give more detailed stats than Webalizer when it comes to analysing your server log - more accurate as well as far as I can tell.

Anything working off the server log has an advantage over stats like Google Analytics in that the server log includes everyone who got the page from the server whereas Google Analytics ignores everyone without JavaScript (typically around 6 to 10% of your visitors).

I agree, Awstats is is fine for most websites.

I’ve been doing some research on GeoIPs. The one that looks set up the best is the GeoLite databases from MaxMind. Has anyone used these? Any comments, raves, complaints?

That is the best solution I found, and the one I use. The only ‘problem’ with the Free one is it (Certainly used to) show AOL users as ALL being American (including UK based AOL users) - this may have changed recently as AOL has been bought out here in the UK.

Sounds good. Most of the packages seem to use requests from a particular IP as the definition of a distinct user. Does that sound right to you? I’ve seen some people argue that it still overcounts and others argue that it undercounts. And Tim’s page (below) suggests that you can use JS script to more clearly identify distinct users and even track where they go. As you point out, that’s going to short-change the stats a bit, but if I can get more accurate data on 90-94% users, it doesn’t seem “too dangerous” to extrapolate the remaining requests based on that. Or does it? And thanks so much for the first-hand experience, Stephen.

Interesting article. The article makes it sound like you combine server side logs and javascript together in your analytics solution. One, you didn’t talk about the geoip lookup which is interesting. I will google that now that I know what keywords to use ( THANKS!:smiley: ) do you have any geoip advice to add to what I’ll find? (In fact, I would think JS would allow you to measure the screen resolution; GA is JS-based, maybe that’s how you pull that part off?)

I code my own analytic application… I prefer log on my own server and what my application grabs are real time than Google or any other application on the web… I’ve compared my log to Google’s and for some reason google ends up being short or some information is just semi different. I know 100% fact that my application works…

Mine can get user’s ip, geo-location (city, state, zip), how many pages, what pages, session, how many times they came back, where they came from etc…

I use a custom log processor I’ve written myself - it does handle location (by doing a geoip lookup on the IP address) - it can’t do things like screen resolution, I wrote an article on the various logging types http://blog.jellymedia.com/2010/02/05/understanding-web-stats/ which you may find useful in explaining the differences between log / analytics methods.

I haven’t tried either of these, I tend to process the log files myself, quick search for ‘open source analytics’ gave me them.

http://www.openwebanalytics.com/