I am working in a startup and we need to setup some analytics / data-browsing so non-tech users can browse the data and produce/save some reports.
The database is running MongoDB and I'd like to avoid ETL techniques that would require a lot of maintenance.
Ideally I would also need a software that lets me write custom code to get complicated variables (eg. using MongoDB aggregate/map-reduce framework)
So the somewhat flexible requirements in bullet points
- Can plug in easily on MongoDB database
- Mostly for static/batch analysis : operation on already existing data
- Can configure and save reports (eg generated the same report with
nowas time reference)
- (Bonus if it also handles events / real-time data like Mixpanel)
- No or very little maintenance required when data changes (new fields in JSON documents)
- Can be used easily by non-tech users (who are only good at using Excel).
- Supports custom code so when special information/variables are needed, a developer can write a piece of code to generate the missing variable
- Price affordable for an early startup (we're talking about 100-200euro/mth). We can afford to have the service running only a few days per week (in a cloud-like fashion) in case that helps.
These are not really hard-requirements, and if one of the app is very good on most of them, then it's worth investigating.
So far we had a look at
- SlamData - I really disliked the UX and at first sight it was hard to really do something you want with that...
- Tableau - Seems to require ETL conversion and opening data with PostgreSQL
- Databazel - was having problems opening a mongoDB connection, getting some weird undocumented quasar error (since SlamData also uses quasar and worked fine it kinda puzzles me)
- Pentaho - Really too expensive
- Softs like Mixpanel only seem to handle event-based information and are not made for batch analysis on already existing data