The Jupyter notebook in this project receives a Whatsapp Chat.txt file and conducts a descriptive and natural language processing analysis on it. The features include:
- The evolution of messaging habits in the conversation, with a break-down per sender
- Wordcloud of most frequently used words in the chat as a whole and per sender
- Sentiment analysis of the texts
- Activation times of the texts
While my original purpose was to create a code for group-chat analysis, it also works for conversations with just one person.
I have provided a test chat file for you to try the code out. If you prefer, you can run the code with your own conversations! The test file is randomly generated, so there should not be any clearly defined patterns. Down below I provide the steps to export your own chat from Whatsapp.
You can use this code to analyze conversations in languages other than English. Just keep in mind that you will have to change the language in the stopwords line from english to the language you desire.
To work with your own data, first you have to open Whatsapp. Then go to the desired chat and click on the three dots in the upper-right corner. It should look like this:
Afterwards, click on More to find the following display:
Choose your desired way of sending the data to yourself and voilá!