-
Notifications
You must be signed in to change notification settings - Fork 28
Home
The Pega Data Science tools help to analyze the value and analytical performance of applications based on Pega AI, machine learning and decisioning, such as implementations of Pega Customer Decision Hub or Pega Process AI.
The tooling consists of scripts, applications and libraries and is Open Source, not an official Pega product. We release this under the Apache 2.0 license. Pega does not make any representation or warranty with respect to this free software. It targets data scientists as well as people with a less specialized role looking more for generic insights. The tooling is all in Python currently. We have a legacy version in R that is no longer maintained, contact the authors if you want access.
Functionality includes
- A standardized ADM Health Check to get insights into the models and predictors. This Health Check can be done from a data science environment but can also, conveniently, be done from a stand-alone application. This application can generate stand-alone HTML ADM overview reports, as well as detailed reports for individual ADM models.
- Examples/starter scripts run your own, custom, ADM datamart analysis.
- Example of a Value Finder analysis that supplements the built-in product features.
- Example analysis of Interaction History.
- Explanatory articles on ADM, Thompson Sampling etc with working examples
- For the more technical, data science oriented users there are methods and classes in Python to work with Pega data, visualize it and more.
- A collection of sample plots/graphs for analysis of CDH, more as inspiration.
Whilst the examples are CDH oriented, the code will generalize to analyzing other Pega AI applications, such as applications based on Process AI.
The documentation contains the API reference as well as the examples and tutorials.
To install the package and get started head over to Installation Instructions
See Contributing