Question

I’ve started a django project that will include an analytics app. I want that app to use either couchDB or mongoDB for storing data.

The initial idea was (since the client already is using Google Analytics) to once a day/week/month grab data from GA, and store store it locally as values in database. Which would ultimately build a database of entries – one entry per user per month – with summed values like

{"date":"11.2011""clicks": 21, "pageviews": 40, "n": n}, 

for premium users there could be one entry per user per week or even day.

The question would be:

grab analytics from GA, do a sum entries for clicks, visits etc.

or

store clicks and whatever values locally and once a month do sums for display ?

Answer 1

Right, old question but I’ve just finished the project so I’ll just write what I did.

Since I didn’t need concurrency and wanted more speed approach, I found that mongodb is better for that.

The final document schema that I’ve used is

{'date': '11.2009', 'pageviews': 40, 'clicks': 13, 'otherdata': 'that i can use as filters'}

The scope of my local analytics is monthly, so I create one entry in mongdb per user per month, and update it each day. As said just now, I update data daily, and store only summaries and averages of those.

What else. Re: Jamie’s answer… The system is using GA events, so I’ve got access to all data that i need.

Hope someone may find it interesting.
cheers and thanks for ideas !