Pythia is a set of tools that work with REMP and CRM data and allows to:
- Predict probability of users conversion and expose this information by creating segments (of low/high conversion probability) for other parts of the system (Beam/Campaign/CRM).
- Predict probability of users churn and store this information in BigQuery database (which can be later used by other tools, e.g. CRM).
Pythia works with different data sources - all of these are mandatory at this point:
- Beam pageviews data (stored in Elastic via JavaScript snippet)
- CRM payments and subscriptions data (stored in MySQL of CRM)
- Conversion/churn model is computed on aggregated data exported to BigQuery (Google Cloud service)
The whole process is split to several steps:
- Aggregation of data and export to BigQuery
- Conversion model training and prediction
- Churn model training and prediction
- Exporting generated data as segments for other applications via API
For installation and running instructions, please see the INSTALLATION file and README files of the linked tools.
To run Pythia tools, you'll need python3 and BigQuery.