Project: Mobile Game Analytics

Overview

This repository contains a complete analytics project based on data from a mobile game.
The project focuses on three key tasks: calculating user retention, analyzing an A/B test of promotional offers, and designing metrics to evaluate an in-game themed event.

The goal is to demonstrate a full analytical workflow: data loading and preparation, metric calculation, application of statistical methods, and business-oriented interpretation of results.

Project tasks

1. Retention calculation
   A Python function was implemented to calculate player retention by days since registration.

2. A/B test analysis
   Two sets of promotional offers were compared to determine the more effective option based on product and statistical metrics.

3. Event metrics design
   Metrics were proposed to evaluate the results of the latest themed event in the game.

Data

The project uses datasets described in the notebook:

• problem1-reg_data.csv — user registration timestamps.
• problem1-auth_data.csv — user login timestamps.
• A/B test data with user group assignment and revenue.

These datasets are used to analyze user behavior, retention, monetization, and the impact of product experiments.

Tech stack

The project is implemented in Python with the following libraries:

• pandas
• NumPy
• Matplotlib
• SciPy
• Pingouin

Approach (workflow)

At a high level, the work was structured in the following steps:

1. Data loading and preprocessing
   Imported registration and login datasets, aligned timestamp formats, and prepared user-level activity tables.

2. Retention function implementation
   Built a function that calculates day‑by‑day retention from the registration date, tested it on the full dataset and on samples, and visualized the retention curve.

3. A/B test evaluation
   Aggregated key product metrics by test group, checked statistical assumptions, selected appropriate tests, and evaluated whether differences between groups are statistically significant and practically meaningful.

4. Event metrics design
   Identified meaningful behavioral and monetary indicators for the themed event, defined metric formulas, and outlined how they could be used to monitor event performance.

Results

By the end of the project:

• A reusable retention calculation function with supporting visualizations was implemented.
• The A/B test of promotional offers was evaluated with statistically justified conclusions about which variant performs better.
• A set of metrics for assessing the impact of in‑game events on user behavior and revenue was formulated.

These results can be used as a template for analyzing other mobile games, structuring product experiments, and monitoring event performance.

Skills demonstrated

The project showcases the following skills:

• calculation and interpretation of retention metrics;
• working with cohort logic and user lifetimes;
• analysis of A/B tests;
• selection and interpretation of product metrics;
• statistical thinking and hypothesis testing;
• analytics in Python;
• visualization of analytical results.

Attribution

This project is based on educational assignments from the "Аналитик данных" course at karpov.courses.

В соответствии с п. 2.1.10 Оферты karpov.courses учебные задания использованы для формирования портфолио при соблюдении условий указания авторства ООО «Карпов Курсы» и размещения ссылки на сайт karpov.courses. В репозитории отсутствует информация, связанная с доступом к базе данных Курса.