Reboot: At Home :Bloomberg Chat processor

I recently attended the Lloyds Reboot #Reboot events, which were fantastic! Although I couldn’t participate as a hacker due to various reasons, I decided to take on a similar challenge at home. Many traders face the issue of analyzing Bloomberg chat #Bloomberg, and I aimed to tackle this problem.

This time, I had a new teammate in my hacker team: #Copilot, Microsoft Copilot.

Problem Statement:

Traders use Bloomberg chat to communicate and close deals. They have a specific way of communicating in these large group chats. Missing a single message can mean missing an opportunity. The goal of this project is to analyze Bloomberg chat and provide real-time data to traders, allowing them to click and find on-demand information about bonds, rates, dates, and actual messages.

Solution:

As a C++ developer, I decided to make things easier by using Python, which is like a Lego set—find a library and connect it. Assuming #BloombergChat provides a chat API to capture messages, I replicated this by using a text file where chat messages can be copied as they arrive.

I started a Python Flask server with HTML5 websockets and multithreading to continuously read new files as they arrive. Thanks to Copilot, I got the correct syntax to insert and update data into an SQLite3 database.

Everything runs in real-time, so when a new chat message comes in, it processes and updates the chart. Websockets push the updates to the trader’s screen, so no one needs to click anything to get new updates.

Demo:

I’ve uploaded a demo at the link below.

Hope this makes sense to someone!

https://youtu.be/OPoGu7GV8qs

Let me know your thoughts!