tolubass/Immuniwatch_baseline_analysis

Baseline analysis of public perception of immunization across social media platforms in Nigeria # Immuniwatch Baseline Sentiment Analysis Public perception analysis of immunization across Nigerian social media platforms using sentiment analysis and natural language processing. --- ## Table of Contents - Overview - Key Findings - Project Structure - Dataset - Methodology - Usage - Main Recommendations - Contributors - Contact --- ## Overview This project presents a comprehensive baseline sentiment analysis of vaccine-related conversations on Nigerian social media platforms. Over **45,000 posts** were collected and analyzed from Facebook, Twitter/X, Instagram, TikTok, and blogs spanning 2008-2025. ### Objectives - Determine overall sentiment distribution (positive, neutral, negative) toward vaccines - Identify key topics and themes driving vaccine conversations - Examine geographic variations in vaccine sentiment within Nigeria - Detect misinformation-related cues and distrust indicators - Establish temporal patterns in vaccine discourse --- ## Key Findings - **Positive Sentiment**: 56.6% (17,756 posts) - **Negative Sentiment**: 40.3% (12,644 posts) - **Neutral Sentiment**: 3.5% (1,103 posts) **Geographic Focus**: Analysis concentrated in Minna (Niger) and Ikeja (Lagos), accounting for >95% of tagged posts. **Platform Distribution**: Facebook (60%) and Twitter/X (39%) dominate vaccine discourse. **Emotional Landscape**: Neutral tone (51.2%), Joy (10.2%), Sadness (8.0%), Anger (7.1%) --- ## Project Structure ``` Immuniwatch_baseline_analysis/ │ ├── data/ # Dataset files │ ├── clean_scraped_posts.csv # Cleaned dataset │ ├── merged_vaccines_data.csv # Merged data │ └── preprocessed_final.csv # Final preprocessed data │ ├── notebooks/ # Jupyter notebooks │ ├── EDA_exploration.ipynb # Exploratory data analysis │ ├── data_cleaning_and_preprocessing.ipynb │ ├── data_inspection.ipynb # Initial data inspection │ ├── sentime …