FinHub Application Demo

FinHub is a financial web application that provides news, analysis, and tools for investors and traders. It offers features such as news aggregation, financial data visualization, and AI-powered insights.

Note: This is only a demo, only deployment can realize its intended potential. We apologize for the performance and other issues,because of not being deployed yet, the computations are all on the client side. With future deployment, the computations will be done on the server side and the issues and perforamnce will be fixed.

Home Page

The main landing page of the FinHub application. It displays a dynamically generated headline image that summarizes the top news topics using Stable Diffusion. The page also provides an overview of the current market sentiment using intuitive data visualizations.

News Page

News aggregation and analysis capabilities of FinHub. Users can browse through a curated list of financial news articles, view summaries, and access sentiment analysis insights generated by the application's machine learning models.

Too Lazy Page

The "TooLazy" feature, which allows users to input video links related to financial content. FinHub then transcribes the audio using OpenAI's Whisper model, summarizes the content with BART, and performs sentiment analysis and keyword extraction, providing a comprehensive overview of the video.

Setup (Please do before running!)

1. Create Venv, install requirements, and activate

.venv\Scripts\activate
cd backend
pip install -r requirements.txt
.venv\Scripts\activate

2. Create .env file in the root directory, set environment keys

OPENAI_API_KEY = paste-my-key-here
STABILITY_KEY = paste-my-key-here
NEWS_API_KEY = paste-my-key-here

3. ffmpeg install (for link-to-audio)

Windows

• Download from https://www.gyan.dev/ffmpeg/builds/ffmpeg-git-full.7z
• Move the ffmpeg.exe and ffprobe.exe to backend/whisper_client/ffmpeg/

Mac (Please contact if problem, not tested since no MAC machine)

• brew install ffmpeg

4. Download our finetuned BERT model

• https://drive.google.com/drive/folders/1f73MSmtG9qHJW4UkEI_FxGBl9dYQqUye?usp=drive_link
• Put the content in backend\bert\model

Start Running!

Backend

In the /backend directory, with venv activated:

python app.py

Frontend

1. npm install
2. npm start
3. navigate to Top News Page (to fetch and compute the news) and then you can proceed to home page. (Apologies for the extra time: because of not being deployed yet, the computations are all on the client side. With future deployment, the computations will be done on the server side and the issues and perforamnce will be fixed.)