Retentioneering Tools Versions Save

Overview

Retentioneering 3.0.0 is now available! This release features significant updates and improvements compared to its predecessors. It includes major changes to the library's architecture and functionality.

You can get started with the new features by referring to the Getting Started section in our documentation

Features and enhancements

• Retentioreering's functionality has undergone a major overhaul, and as a result, it no longer operates as an extension of pandas.DataFrame. Instead, a class system has been implemented to improve the library's architecture and functionality.
• We’re excited to announce the addition of a new class to the retentioneering library - the Eventstream class! This powerful data structure is designed to serve as a core component for clickstream data storage, preprocessing, and seamless integration with retentioneering’s analytical tools.
• We have added a set of Data processors user guide to the library which includes a library of typical operations on eventstream (clickstream), that can be combined into preprocessing scripts. This functionality has significantly expanded the capability of the library and allows users to preprocess their data in a more efficient and customizable manner.
• Two new tools have been added to Retentioneering in version 3.0.0: Step Sankey and Cohorts.
  • Step Sankey allows users to create visualizations of user flows through various steps or events in a process.
    • User guide
    • API references
  • Cohorts enables users to analyze user behavior and retention rates based on cohort segments.
    • User guide
    • API references
• We have added new features and improvements to the existing tools:
  • Transition Graph. We have made significant changes to the Transition Graph tool to provide a more intuitive and user-friendly visualization of user flow across different stages of the user journey.
    • User guide
    • API references
  • Funnel. Now includes two new types of funnels, providing more flexibility in analyzing conversion rates.
    • User guide
    • API references
  • Clusters. We have expanded the functionality of the Clusters tool by adding two types of vectorization, removing restrictions on the number of available colors, and enabling the ability to set your own clustering methods.
    • User guide
    • API references
  • Stattests (ex. Compare groups). We have added new statistical tests to the Compare Groups function, providing more options for comparing user groups.
    • User guide
    • API references
• The step matrix tool has been successfully ported to the new architecture of Retentioneering 3.0.0 without any major changes. Users will continue to enjoy the same functionality and benefits of this tool.
  • User guide
  • API references

Changes to syntax and compatibility

• Retentioneering 3.0.0 has changed the syntax of working with the library. It now uses a class system instead of an extension of pandas.DataFrame.
• There is no backward compatibility with previous versions of the library.
• Retentioneering 3.0.0 supports Python versions 3.8, 3.9, and 3.10, Python 3.7 has been dropped.

Documentation

• The official documentation for Retentioneering version 3 is available at https://doc.retentioneering.com/3.0/doc
• We have added tutorials to our documentation that demonstrate how to use our instruments for various tasks.
• We have also included descriptions of the datasets we use in our documentation and tutorials.

Optimized environments for visual tools

To ensure the best performance when using our visual tools, we recommend running them in the following environments:

• Environment
  • Jupyter Notebook
  • Jupyter Lab
  • Google Colab
• Browser
  • Google Chrome

Completely reworked Retentioneering workflow: functions init_config() and retention.prepare() are removed. In 2.0 it is not required to initialize “positive” and “negative” events before the analysis. You can start exploring your user behavior data first and define targets when needed as optional parameters. To access all Retentioneering tools attribute “.retention” was renamed to “.rete”. To get started with an updated workflow refer to this guide.

plot_step_matrix() function was significantly reworked and renamed to step_matrix(). To read more about new step_matrix() functionality refer to this description

get_edgelist(), get_adjacency() and plot_graph() functions now have customizable weighting options (total nuber of events, normalized by full dataset, normalized by nodes, etc.). To learn more please fere to this description

new function compare() was added to compare two segments of users or test/control groups in AB test based on defined metrics. Read more about compare function in this description

Users' behavior segmentation was reworked and now works significantly faster and updated with new functionality. For more information refer to this description

Tools to plot user conversion funnels were reworked and improved. To learn more about new features read this description