I'm surprised BitYoga chose MongoDB for real-time analytics. Several years ago we attempted to do a real-time analytics solution with MongoDB but besides being a not so great performer when it comes to counting, it's boolean operators were still in its infancy. We ended up ripping out and replacing with another back-end solution in a couple of months and never looked back. Has MongoDB changed much to make real-time more realistic?
Hi Dishwasha We don't use Mongo for analytics - it is an upstream store for transactional data for us. We are a Data Warehouse Service for Analytics and we enable fast SQL-based analytics in our system for data from Mongo and other JSON (semi-structured ) sources its fast because (a) you don't need to do ETL on your Mongo/JSON data before analysis - so you save that time and the temporal value of fresh data is preserved (b) we have a scale-out MPP architecture to add compute and storage as needed check us out at http://www.bityota.com/