User Activity Analysis

Analysis of customer behavior base on statistics from online shop site (public datased from Kaggle).

Dataset: eCommerce behavior data from multi category store

Requirements: requirements.in

How to set up the data

Run the following scripts:

python upload_data.py
python missing_values_imputation.py
python anomaly_removal.py
python create_additional_tables.py

It will upload data to the local DuckDB database, which will be created automatically. It will also clean the data and create some additional tables for the analysis.

Analysis

Plan

1. Data preparation and EDA

   • Upload data to DuckDB for storage.
   • Handle missing values.
   • Remove anomalies.

2. [WIP ]Describe user behavior.

   • User Funnel with conversions.
   • User Retention and cohorts.
   • User LTV.

3. [WIP] User Segmentation. Come up with different user features, which can be used for further analysis.

   • Behavioral segmentation.
   • Segmentation by purchase.

4. [WIP] Insights searching. Find the potential customer problems. First of all, chose metric for measuring user success. Then compare segments of users by success metrics and find weak segments, or compare to some benchmark of the industry.

Tech stack

• Python libraries:
  • Data Wrangling: numpy, pandas
  • Data Visualization: seaborn, plotly
  • pip-tools for human-readable pinning of dependencies.
• Jupyter Notebook for interactive analysis.
  • JupySQL plugin for SQL queries
• DuckDB - for data storage and fast analytical computations.