whatsapp-chat-analyze 0.1.7

Ingest and analyze WhatsApp chat data, and plot beautiful visualizations.

whatsapp-chat-analyze

Command line tool to analyze WhatsApp exported chat data (accepts .txt and .zip) and plot pretty interactive charts.

• whatsapp-chat-analyze
  • Demo
  • Features
  • Installation
    • pipx
    • pip
  • Usage
  • Develop
  • Limitations

Demo

For more plots and interactity, check out the blog post: https://teddysc.me/blog/whatsapp-chat-analyze .

Features

• Simple to install and use, no cloning involved, supports modern Python versions
• Extracts chat data from .txt or .zip files
• Export to csv (-c)
• Pretty interactive charts with plotly and save them to HTML files
• 6 different plots, see demo
• Anonymize sender names to A, B, C, etc. (-a)

Installation

Python>=3.10 required.

pipx

This is the recommended installation method.

$ pipx install whatsapp-chat-analyze

pip

$ pip install whatsapp-chat-analyze

Usage

$ whatsapp-chat-analyze --help

usage: whatsapp-chat-analyze [-h] [-n name] [-o base] [-d] [-E] [-c] [-a] file

Analyze Whatsapp Exported .txt or .zip (will be automatically extracted) chat file

positional arguments:
  file                  Chat file (_chat.txt or *.zip) to analyze

options:
  -h, --help            show this help message and exit
  -n name, --chat-name name
                        Name of the chat (default: Chat)
  -o base, --output-base-name base
                        Output base name for the plots (default: whatsapp-chat)
  -d, --by-day-only     Plot messages per day only (default: False)
  -E, --extract-only    Extract the chat and exit (default: False)
  -c, --to-csv-only     Convert chat to csv and exit (default: False)
  -a, --anonymize       Anonymize the chat by replacing author names with generic names (default: False)

Develop

$ git clone https://github.com/tddschn/whatsapp-chat-analyze.git
$ cd whatsapp-chat-analyze
$ poetry install

Limitations

• Parsing of chat data uses simple regex, and author names in non-latin languages may not be parsed correctly