twitch-chat-analyzer 0.0.3

Twitch chat analyzer from past broadcasts, designed for Jupyter notebook

twitch-chat-analyzer-py

Twitch chat analyzer from past broadcasts, designed for Jupyter notebook.

Getting Started

Installation

The package can be installed by pip command in terminal

pip install --upgrade twitch-chat-analyzer

Run Jupyter Notebook

The package is mainly intended to be used in Jupyter notebook, but it can still be used in standard Python environment

Instruction to install Jupyter notebook (or newer JupyterLab) is here

After installation, run the following command in terminal to start Jupyter

jupyter notebook or jupyter-lab

Download chats from past broadcast

notebook.ipynb has examples of statistics functions.

from twitch_chat_analyzer import analyzer

# Create an analyzer object from video ID.
# If the chat log was not downloaded before, it will download automatically and create an analyzer.
ann = analyzer.FromVideoId('REPLACE_HERE_TO_VIDEO_ID')

# Some pre-built statistics functions to draw graph
ann.DrawChatPerMinutes(10)  # Chat counts for each 10-minute interval
ann.DrawTopChatters(20)  # Top 20 viewers with most chats
ann.DrawTopEmotes(15)  # Top 15 most used emotes

# If you want to handle dataframe yourself
df = ann.ToDataFrame()

DataFrame

The dataframe returned from ToDataFrame() has the following columns

Column name	type	meaning
offset	float	Time of the chat, in seconds after stream started
username	str	Twitch login username
display_name	str	Display name in chat, which may not be in English
name	str	Combined name of username and display_name, as displayed in Twitch chat
body	str	Raw chat content, including emote text
text_body	str	Chat content excluding emotes
is_subscriber	bool	if the chatter is a subscriber
bits	int	Amount of bits spent in the chat
is_sub_notice	bool	if the chat is new/renew subscription notice