While working at onefinestay, I developed the company’s data analytics infrastructure.
There were two main components - pylytics, which is an ETL (extract, transform, load) tool, and Saiku, which was used to visualise the data.
Pylytics was developed in house, and is open source.
The purpose of pylytics is to take data from a range of sources, clean it, and then create a star schema.
Below is an example of a simple star schema. The sales ‘fact’ points to two ‘dimensions’ (date and location). Even this simple example allows powerful queries, such as addregating sales by country, city, day, month, year etc.
Most star schemas have many more dimensions that this, which allows for powerful analysis.
Saiku is a tool which can be used to analyse the data in a star schema.
Here is an example graph which Saiku can create:
Here are some screenshots of Saiku: