README.md

metadata

title: Nifty News Analysis
emoji: 📈
colorFrom: blue
colorTo: green
sdk: streamlit
sdk_version: 1.45.1
app_file: app.py
pinned: false
license: mit

Nifty News Analysis

A Streamlit app that analyzes NIFTY 50 stocks and news, built by @MtotoWaJemo.

NIFTY 50 News Analysis

Overview

This project is a Streamlit-based web application that analyzes news sentiment for companies in the NIFTY 50 index, categorized by sectors. It fetches news articles using the NewsAPI, summarizes them using the T5 model, and performs sentiment analysis using DistilBERT. The app provides insights into sector and company sentiment to guide investment decisions.

Features

• Sector Selection: Choose from NIFTY 50 sectors (e.g., Financials, Healthcare).
• Time Frame Analysis: Analyze news from different periods (1D, 5D, 1M, 6M, YTD, 1Y, 5Y).
• Sentiment Analysis: Summarizes news and classifies sentiment as Positive, Negative, or Neutral.
• Investment Insights: Provides sentiment scores and recommendations for companies.
• Interactive UI: Built with Streamlit, featuring a user-friendly interface with tables and visualizations.

Installation

1. Clone the repository:

   git clone https://github.com/mtotowajemo0/nifty-news-analysis.git
   

2. Navigate to the project directory:

   cd nifty-news-analysis
   

3. Install dependencies:

   pip install -r requirements.txt
   

Requirements

• Python 3.8+
• Libraries listed in requirements.txt:
  • streamlit
  • newsapi-python
  • transformers
  • streamlit-extras
  • pandas

Usage

1. Obtain a NewsAPI key from newsapi.org.
2. Replace the api_key in app.py with your NewsAPI key.
3. Run the Streamlit app:

   streamlit run app.py
   

4. Open the app in your browser, select a sector and time frame, and click "Analyze News" to view results.

Files

• app.py: Main application script with Streamlit code, news fetching, and sentiment analysis.
• requirements.txt: List of Python dependencies.
• README.md: Project documentation (this file).

Notes

• The app uses the t5-small model for summarization and distilbert-base-uncased-finetuned-sst-2-english for sentiment analysis.
• News articles are filtered based on a weighted keyword system to ensure relevance.
• Sentiment scores are calculated as (Positive - Negative) / Total Articles.
• Disclaimer: Insights are for informational purposes only and not financial advice.

License

MIT License. See LICENSE for details.