We created a simple API which gives us data ranges such as
- /:start/:end - number of users registered within date range
- /:start/:end/active - number of users registered within date range who were active within last week
etc. etc.
Then it's pretty straightforward to use ImportData & CONCATENATE, as well as fill down to create a matrix of cohorts, which can in turn be graphed.
In general it's best to pull in the raw data with ImportData and then calculate percentages and such in the spreadsheet. The latter are updated when the cells they reference refresh.
Docs refreshes ImportData sources once every hour, caching them in between, so the load on the server and DB is minimal and you end up with more or less current stats.
- /:start/:end - number of users registered within date range
- /:start/:end/active - number of users registered within date range who were active within last week
etc. etc.
Then it's pretty straightforward to use ImportData & CONCATENATE, as well as fill down to create a matrix of cohorts, which can in turn be graphed.
In general it's best to pull in the raw data with ImportData and then calculate percentages and such in the spreadsheet. The latter are updated when the cells they reference refresh.
Docs refreshes ImportData sources once every hour, caching them in between, so the load on the server and DB is minimal and you end up with more or less current stats.