A deep learning project to find trending news in African countries
# What's The Hot Topic in Topic Town?
## Information Retrieval & Natural Language Processing
A deep learning project to find trending news in African countries
**Contributors:** Kelvin K. Ahiakpor & Emmanuel A. Acquaye
**Institution:** Ashesi University
**Contact:** kelvin.ahiakpor@ashesi.edu.gh, emmanuel.acquaye@ashesi.edu.gh
### Overview
This project addresses the rampant spread of misinformation and disinformation through credible news reporting using Natural Language Processing (NLP) and Hugging Face Transformers. It aims to provide verified, insightful, and tailored information by analyzing news and understanding underlying sentiments.
This project automates the extraction, summarization from "All Africa", storing all scraped news data in an SQLLite3 database. It is designed to provide on-the-go accessible news reporting to individuals, journalists, and social trend enthusiasts with trending news topics in African countries.
### Application Demo
Link to application demo: App Video Demo
### Interactive app link
Link to interactive app: Interactive App
### How to run
1. Ensure a strong and stable internet connection, you may follow the setup first for local hosting here
2. Click the link to open the Interactive App. If the app is asleep, gently wake it up, otherwise reboot the app!
3. Navigate to the drop down list, click `SEARCH` and type your African country
4. Wait about 2 minutes to get news, sentiment analysis and summaries.
5. Note, the more news your country has, the more likely it is that we will have to process more data.
6. If scraping is taking a while, please hang on.
7. Thank you
### Setup
1. Ensure you have the following libraries installed:
`tensorflow`
`opencv-python`
`pandas`
`numpy`
`streamlit`
`scrapy`
`pycountry`
`beautifulsoup4`
`transformers`
`scikit-learn`
`torch`
`evaluate`
`accelerate`
You may use the following command:
`!pip install tensorflow opencv-python pandas numpy streamlit scrapy pycountry beautifulsoup4 transformers sciki …