Inspiration

To build a tool that can be easily used by discord or any other social media platform channel's owner to monitor real time sentiment of the chats.

What it does

The user can use this tool to analyse and monitor the channel's chat real time,like the sentiment of the chats and filter out a specific message based on it's sentiment value or toxicity

How we built it

I built it using Django, a python web framerwork and using the API provided by ExpertAi for this hackathon and using Discordpy to read messages and send it back to the server for further sentiment analysis and other operations. Dashboard is taken from creative-tim and modified specific for this project to show real time value.

Challenges we ran into

First,using Discordpy library might not be the right choice to integrate with the current project, since we can't send messages directly to the server, so instead we have to write in a text file everytime we recieve a new message and then the file is read back in the backend of Django. Secondly, The Dashboard still miss some of the functionalities like easy integration with chats and the graphs scale could be improved too.

Accomplishments that we're proud of

I have successfully demonstrated that we can use it to see real time sentiment of the chats by showing it on charts and can filter out messages with a specific sentiment values.

What we learned

Learned about ExpertAi as NLAPI and made my first discord bot for this project. Also, On frontend side, I learned a little bit about JavaScript and Ajax and how to use them for better integration of frontend and backend.

What's next for Discord Chats Analysis

The Project could have many potential usecases like identifying the toxicity level in the chat and can be used to capture useful informations by analysing each message. The major usecase could be that instead of relying on a bot that is not always perfect to moderate the chat, this tool can be use to get an intital idea about the vibe of the chats and filter out users and messages according to the toxicity which is far more easy to verify among thousands of messages and users in the chat.

Built With

Share this project:

Updates