Hello Soylentils. I work in the IT Department of a mid sized utility company. Our engineering department requested that we use Google Analytics to track how people are hitting certain pages, from where, and whether or not they are getting to pages from links on our website or directly. A co worker found that Google Analytics can get expensive. Does anyone have any experience with any FOSS alternatives such as the ones listed in the article below?
http://opensource.com/business/14/10/top-3-open-source-alternatives-google-analytics
This discussion has been archived.
No new comments can be posted.
Ask Soylent: FOSS Equivalent to "Google Analytics"
|
Log In/Create an Account
| Top
| 30 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(Score: 2, Touché) by Ethanol-fueled on Tuesday August 04 2015, @05:35PM
Do what every other utility does and just pass that cost on to the ratepayers. Hell, do utility companies even need excuses to keep jacking up their rates year after year?
Problem solved.
(Score: 3, Informative) by tibman on Tuesday August 04 2015, @05:38PM
I could swear SN use to have piwik running? *Searches* Yup, here is NCommander's post on it: https://soylentnews.org/article.pl?sid=14/06/06/1923229 [soylentnews.org]
It doesn't appear to be running anymore though.
SN won't survive on lurkers alone. Write comments.
(Score: 1, Informative) by Anonymous Coward on Tuesday August 04 2015, @06:43PM
We've been using piwik for our sites for several years. Haven't had any issues with it, plus it works in real-time. The data tends to build up, but the software is very complete and very responsive when viewing reports/stats/whathaveyou.
(Score: 3, Informative) by NCommander on Tuesday August 04 2015, @11:55PM
It was only temporary, and we took it down after we collected the data we wanted. It may reappear in the future after we give heads up to the community.
Still always moving
(Score: 2) by TheRaven on Wednesday August 05 2015, @08:59AM
sudo mod me up
(Score: 2, Informative) by opinionated_science on Tuesday August 04 2015, @05:42PM
Isn't this just apache logs and grep?
(Score: 3, Funny) by tibman on Tuesday August 04 2015, @05:51PM
Only if you like reports with zero aggregation : ) Which sounds like an oxymoron.
SN won't survive on lurkers alone. Write comments.
(Score: 4, Informative) by Lunix Nutcase on Tuesday August 04 2015, @06:22PM
No, it's not. Not even close. Have you ever used Google Analytics even once?
(Score: 0) by Anonymous Coward on Tuesday August 04 2015, @06:29PM
What data do they collect that you wouldn't find in Apache logs?
(Score: 3, Interesting) by pendorbound on Tuesday August 04 2015, @07:11PM
In terms of what they *collect*, lots of fingerprinted info about the browser. Versions, plugins, screen resolution, etc. More than just the user agent string in a combined log contains because GA executes JavaScript code on the client to do additional detection.
Beyond the collection though, the reporting they provide is substantially above what you'd be likely to cook up with a shell pipeline. Even if you completely eschewed graphs (their graphs are very useful) and said tables of data were good enough, you'd be hard pressed to present the data in as many useful forms as GA does.
Came here mostly to say piwik though. I've been using it for a couple of years, and it's a pretty good approximation of what GA gives you, without you giving Google anything more than they already get by way of your users' search results & DNS queries.
(Score: 2) by Lunix Nutcase on Tuesday August 04 2015, @08:10PM
It's not necessarily just the data but also the aggregation and visualization.
(Score: 2) by opinionated_science on Tuesday August 04 2015, @07:45PM
Nope. Why I asked...
(Score: 2) by TheRaven on Wednesday August 05 2015, @09:01AM
sudo mod me up
(Score: 2) by DarkMorph on Tuesday August 04 2015, @09:34PM
(Score: 2) by Yog-Yogguth on Friday August 07 2015, @12:56AM
Same here, I don't allow it anywhere and don't think anybody should. By default I block everything.
Bite harder Ouroboros, bite! tails.boum.org/ linux USB CD secure desktop IRC *crypt tor (not endorsements (XKeyScore))
(Score: 0) by Anonymous Coward on Tuesday August 04 2015, @05:56PM
Does any of the alternatives run locally (that is, does not require contacting a server owned by the provider of the tool or another third person)?
(Score: 0) by Anonymous Coward on Tuesday August 04 2015, @07:07PM
webalizer [webalizer.org]
(Score: 1, Informative) by Anonymous Coward on Wednesday August 05 2015, @05:33PM
You instal piwik on your own server.
(Score: 0) by Anonymous Coward on Tuesday August 04 2015, @06:19PM
Why is it that you absolutely never reach for the answer of creating it yourself?
I'm astounded at how crippled everyone thinks they are.
First understand that all analytics is, is an X,Y on a chart. So at your server side take a look at what google is storing for X,Y and set your system up to store that number.
This would be a server side operation, I recommend nodeJs for this as it offers easy subtle control timers etc etc.
Then you make an interface, you more or less only make 1 because they only do graph charts anyways. You could toss in pie charts, but again add an extra day for that.
How long does it takes you? About 3 days.
Should this be the end of the world such that your absolutely incapable of just handling this problem like a man by yourself? No.
(Score: -1, Flamebait) by Anonymous Coward on Tuesday August 04 2015, @06:45PM
Yo, douche nozzle, I bet you're used to "handling it yourself" if you know what I mean.
(Score: 0) by Anonymous Coward on Tuesday August 04 2015, @06:49PM
I just did this.
I needed a system that would do an HTTP/HTTPS "health check" on a bunch of webservers. Scoured the web etc -- nagios et al were too much for my needs. I looked at some of the free cloud monitoring services and found them to be lacking, unreliable, or inane.
I wrote a ten-line python script that basically did it myself. It checks to see that I get a 200 back from a list of webservers, and emails/texts me with problems.
Never did find a good F/LOSS solution.
(Score: 2) by GungnirSniper on Wednesday August 05 2015, @12:05AM
Did you consider open sourcing your own script? Just put it in the comments here, Anonymous.
Tips for better submissions to help our site grow. [soylentnews.org]
(Score: 2) by hendrikboom on Wednesday August 05 2015, @03:21PM
Don't you get it? Python is your FLOSS solution. Your so-called ten-line script is just the configuration file.
(Score: 2) by GungnirSniper on Tuesday August 04 2015, @07:58PM
No one ever got fired for using GA.
Chances are even if the typically-overworked IT denizens could create something similar from scratch, it will be different from the tools that management already knows and is comfortable with.
Tips for better submissions to help our site grow. [soylentnews.org]
(Score: 1, Insightful) by Anonymous Coward on Tuesday August 04 2015, @08:44PM
No one ever got fired for using GA.
Well, they should. This privacy-invading nonsense that involves third parties like Google, Facebook, etc. should be heavily discouraged. But most web 'developers' are utterly and completely incompetent.
(Score: 0) by Anonymous Coward on Tuesday August 04 2015, @09:42PM
NodeJs? Fanboy alert. KISS approach: Create a tiny borderless IFRAME on your web pages that sends URL parameters to a PHP (or other LAMP) script on another server. Make sure it's a different server so that write bottlenecks don't take down your site. Your script would take these URL parameters and append them to a flat log file, along with IP address and any other HTTP parameters you want to store. You may want to have the script split the log-file into one-day-per-file, using a formatted version of the system date as part of the destination file name.
The URL parameters are whatever you want to track, such as page number, section of web site, etc. (depending on CMS or site conventions.)
Then you can process these log files any way you want to get stats. An experienced MS-Access coder should be able to generate typical reports if you don't want to do that part yourself.
There a few things Google may be able to do that you can't readily get this way, such as translating HTTP browser type string into "friendly" titles, at least not without building or finding your own translation list. Google can also guess location based on IP address, using their huge internal look-up tables. There may be OSS equivalent lists out there.
Because IP addresses are not unique, you may want to consider using session cookies to track visit duration etc., but some users get ticked off by that.
(Score: 0) by Anonymous Coward on Tuesday August 04 2015, @07:06PM
Have a look over at Prism Break: https://prism-break.org/en/categories/servers/#web-analytics [prism-break.org]
(Score: 1) by miljo on Tuesday August 04 2015, @07:36PM
It takes an eye for looking at the network up, but I've been playing around with using sFlow to collect application performance metrics and coupling them with our network monitoring data. A very different approach, which isn't for everyone, but with enough thought and development, I think I can get a reasonable simulacrum of GA using the tools we already have in place.
Some more info here.
http://www.inmon.com/tutorials7/applications.php [inmon.com]
What's more, if you couple this with sFlow-RT, you should be able to see of your stats in real time.
http://www.inmon.com/products/sFlow-RT.php [inmon.com]
One should strive to achieve, not sit in bitter regret.
(Score: 2) by Techwolf on Tuesday August 04 2015, @08:29PM
Just hook it up to apache or nginix logs and get nice reports that google can't get. Like 404 errors and file downloads.
Or set up a analytics server and use the client side javascript bug.
Piwiki even has an android app for viewing reports.
http://piwik.org/ [piwik.org]
(Score: 0) by Anonymous Coward on Thursday August 06 2015, @07:03AM
https://en.wikipedia.org/wiki/Category:Free_web_analytics_software [wikipedia.org]
Anything is better than using the spymaster's offering.