snowfinch

Realtime web analytics

131

Ruby

Snowfinch

Snowfinch is a realtime web analytics application written in Ruby, and
using MongoDB. It provides everything you need to track and visualize
analytics from multiple sites. While it may not be as full featured as
commercial alternatives, it is free and released under the MIT license.

Currently, Snowfinch supports: tracking of pageviews, active visits, and
visitors for a site; pageviews and active visitors for a given page;
monitoring based on a URI query key-value pair, or based on any given
number of referrers (think campaigns or tracking the number of visits
from social media sites). It’s not much, but this is only the beginning.
Take a look at the Roadmap section to see what’s coming.

Getting started

You will need recent versions of Ruby (1.9.2 recommended) and MongoDB.

If you don’t have Bundler installed, install it with:

gem install bundler

Clone the repository:

git clone https://github.com/jcxplorer/snowfinch.git

Edit the following files to suit your needs:

config/database.yml
config/snowfinch.yml

Some information is stored in a relational database, so you can use any
that is supported by Rails. By default it will use PostgreSQL. Edit the
config/database.yml and Gemfile files accordingly if you want to use
a different database.

Install the application dependencies by running bundle in the
application directory.

Run rake db:setup to create a database and an initial user.

You are now ready to launch the application using your favorite Ruby
application server such as Passenger or Unicorn. Sign in with the email
address [email protected] and password snowfinch. Don’t forget to
change those credentials on your account page!

High performance and scaling

Are the pages you are tracking getting too many hits? Good for you!
Fortunately you don’t need to deploy entire instances of Snowfinch in
multiple hosts. All data collection and storing is a done by a Rack
application available as a gem. Just install the snowfinch-collector
gem, and deploy using a config.ru file similar to the following
example:

require "snowfinch/collector"
Snowfinch::Collector.db = Mongo::Connection.new.db("snowfinch")
run Snowfinch::Collector

Now you can deploy that on several hosts and get the web scale fix you
were looking for. Throw some MongoDB shards at it as needed.

By default snowfinch-collector is mounted in the Rails application at
“/collector” for easier deployment. If you run your own instance,
remember to make sure that the MongoDB database is the same that the
Rails application is using, and configure the URI to the collector in
config/snowfinch.yml.

Just running one separate instance of snowfinch-collector will perform
better than when mounted in the Rails application, as there will be no
overhead from the Rails router and middleware.

Roadmap

Display data for any given period of time (currently the past 2 days).
Collect and display referrer information.
Custom visitor tagging.
Geographic location of visitors.
Filter by custom page metadata (e.g. for A/B testing).
Asynchronous data collector.

Reporting bugs

If you think you found a bug, please file an issue on GitHub. Please try
your best to provide steps on how to reproduce the issue you are
experiencing. Failing tests are highly appreciated!

Contributing

If you want to contribute, fork away and send a pull request. If you
want to be sure something will be merged before spending time on it,
feel free to contact me. Don’t worry, hard work won’t be thrown away.

Always try to provide tests with your pull requests. If you’re not sure
on how to test something, mention it on the pull request so that you can
get some help.

I appreciate if you try to stick to the existing coding style, but I can
always refactor later on. If you can stick to 80 characters a line,
please do. If you don’t yet, try it. It makes you write better code.