Geographical Understanding of Twitter Health Topics using NLP
Portfolio project
Geographical Understanding of Twitter Health Topics Using Topic Modeling 🩺🌍
A machine learning-based approach to analyze health-related topics from Twitter data, providing insights into global health trends.
📋 Overview
With the growing prevalence of health-related discussions on Twitter, this project explores the use of Natural Language Processing (NLP) and Topic Modeling Algorithms to identify and analyze trends in global health topics. By leveraging unstructured tweet data, the project extracts and visualizes insights, enabling a better understanding of public health discourse.
🔍 Key Features
Real-Time Data: Extracts live tweets using the Twitter API.
Advanced Preprocessing: Implements techniques like tokenization, stop-word removal, and lemmatization.
Topic Modeling: Uses Latent Dirichlet Allocation (LDA) for clustering and identifying key topics.
Visualization: Generates word clouds and topic distributions for easy interpretation.
Custom Query: Allows users to input location and query terms to extract specific health-related tweet data.
📊 Methodology
Data Collection:
Live tweets are extracted using the Twitter Developer API.
Input parameters: Location and query keywords.
Data Preprocessing:
Stop-word removal, tokenization, and lemmatization.
Conversion of text into numerical vectors using TF-IDF and Count Vectorizer.
Topic Modeling:
Application of the LDA algorithm to identify dominant topics from the tweets.
Visualization:
Word clouds and graphical representations of the identified topics.
📂 Project Structure
Datasets:
No static datasets. Data is collected live from Twitter based on user inputs.
Algorithms:
LDA, TF-IDF, and Count Vectorizer for topic extraction.
Tools and Libraries:
Python, Tweepy, Scikit-learn, Matplotlib, Word Cloud, NLTK.
🛠 System Requirements
Hardware:
OS: Windows/Linux
Processor: Minimum Intel i3
RAM: 4 GB or more
Storage: 256 GB SSD
Software:
Python 3.7+
Anaconda/Jupyter Notebook
Required Python libraries:
tweepy, nltk, scikit-learn, matplotlib, wordcloud📊 Results
Analyzed over 40 major cities for health-related tweet data.
Successfully identified trends and key topics across various health issues.
Provided actionable insights through word clouds and topic distributions.
🌟 Future Enhancements
Incorporate stronger feature extraction methods for higher accuracy.
Build an intuitive user interface (UI) for better interaction.
Use larger datasets for improved topic modeling.
Extend the model for real-time monitoring and trend prediction.
📚 References
Dataset: Generated live using Twitter API.
🤝 Contributors
Parakala Venkata Anirudh - anirudhparakala@gmail.com
Podduturi Hruthvik Reddy
Shaik Sohel
Thunuguntla Srichakranath
📧 Contact
For inquiries or collaborations, contact Parakala Venkata Anirudh at anirudhparakala@gmail.com.
🧐 Fun Fact
Social media is faster than traditional media, making it a rich source of live health-related information!
1 like · 22 views · Aug 10, 2025
Analyzed Twitter data to identify global health trends using NLP and topic modeling.
