andikawilliam / line-chat-analyzer

Chat analyzer for LINE chats using python

Line Chat Analyzer

This is an ongoing repository for the line chat analyzer. The goal of this project is to find out a chat's basic stats. I made this project since LINE doesn't provide stats for your chats (to my knowledge) and I couldn't find anyone who already made an analyzer for LINE chats. It is written in Python making use of pandas and matplotlib for analysis and visualization.

It works by analyzing the chat's history which is provided by LINE on its mobile app. Text data is first cleaned and then converted into a dataframe. From there we can count a user's total number of chats, stickers, shares, etc. Since chat history formats for personal chat and group chats are the same, this code should work for both.

Current Features

• Frequency of chats per hour -> most active hours
• Frequency of chat (including shares, stickers, photos) per user
• Frequency of post shares per user
• Frequency of sticker use per user
• Frequency of photo adding per user
• Frequency of chat (text only) per user
• Avg of chat length by character per user
• Avg of chat length by word per user

Specifications:

• Python 3.7
• Libraries: Pandas, Yaml, wordcloud,

---

How To Use:

Retrieving Chat History

On how to retrieve chat history as text files (https://help.line.me/line/?contentId=20004537)

1. Tap the menu icon at the top of the chat screen > Other settings.
2. Tap Export chat history and choose how you want to send the file.

Before you transfer the text file from your phone to your computer, create a folder with same name as your group name separated by comma. This convention serves to help you in the future since you need to specify the name of the folder as well as the name of the file in the config. Then create two folders named 'cleaned' and 'original. Place the file on the original.

Group_Number_1_chat_history.txt --> folder: Group_Number_1

|--- chat_history
    |--- Group_Number_1
        |--- cleaned
        |--- original > put file here

Analyzing Chat

1. Configure folder and filename on env.yaml

   • for group chats there are cases where users possess multiple accounts. To count their chats as one, specify the duplicated name with the real names

   folder : 'Group_Number_1'
   filename : 'Group_Number_1_chat_history.txt'
   properties :
   duplicate_authors :
       Potato : "Richard Hendriks"

2. On your terminal, run:

   $ python3 main.py
   

3. OPTIONAL: If you only want specific features, simply edit and comment the ones that are not desired.
   NOTE: For now, some features are dependent on the state of the dataframe according the order it was processed. For instance, to get the frequency of chats (text only) per user shares and stickers needs to be deleted.

   Example: Getting text only chat messages

   Edit main.py to only include:

   TextManager = TextManager(cleaned_file_location)
   TextManager.read_file_into_dataframe()
   
   TextManager.delete_non_message_rows()
   TextManager.rename_duplicate_author_identities(
       data['properties']['duplicate_authors']
   )
   
   TextManager.delete_sticker_and_photo_messages()
   TextManager.delete_shares_and_add_photo_messages()
   TextManager.reset_row_index_after_deletion()
   
   TextManager.add_message_length_columns()
   TextManager.get_authors_num_of_messages()

TO DO:

• Add graph generation
• Add testing
• Config menu to choose which stat to retrieve
• Refactor main.py module
• Web implementation