lchuol73/Nigerian-2025-integrated-Vaccination-campaign-sentiment-analysis-during-the-first-7-days

This project performs sentiment analysis on public conversations around Nigeria’s 2025 integrated vaccination campaign to understand how people perceive the campaign, detect misinformation trends, and inform communication strategies for public health stakeholders. The data from X (formerly Twitter), but the same workflow applies to other platforms --- ## 🧑‍💻 Author **Lino Chuol** Health Data Scientist | Digital Health Transformation | Data & AI for Public Health 📍 Nigeria / South Sudan 📧 *github.com --- ## 📝 License This project is licensed under the **MIT License** — see the LICENSE file for details. # 🇳🇬 Nigeria 2025 Integrated Vaccination Campaign — Sentiment Analysis ## 📌 Overview This project performs **sentiment analysis on public conversations** surrounding Nigeria’s 2025 Integrated Vaccination Campaign to: - Understand how people perceive the campaign, - Detect misinformation trends, and - Inform communication strategies for public health stakeholders. The analysis is designed for **reproducibility and clarity**, using clean, modular code that runs cell by cell in a Jupyter Notebook. Although the example uses data from **X (formerly Twitter)**, the workflow can easily be adapted to other sources such as Facebook, news websites, or forums. --- ## 🎯 Key Objectives - 🟢 **Measure overall sentiment** (positive / neutral / negative) about the campaign. - 🧭 **Identify common concerns** and misinformation-related keywords. - 🗺️ **Map sentiment geographically** by Nigerian state (when location data is available). - 📈 **Track sentiment trends** over time (daily/weekly). - 📊 **Produce aggregated, anonymized outputs** suitable for dashboards and reporting. --- ## 🛠️ Project Workflow (Step-by-Step) 1. **Environment Setup & Packages** Install dependencies and prepare the notebook. 2. **Data Collection** Collect public posts via X API or load existing datasets (CSV format). 3. **Preprocessing & Cleaning** Normalize text, remove URLs, mentions, stopwords, and unwanted symbols. 4. **Sentiment Analysis** Use libraries like `VADER`, `TextBlob`, or transformer-based models to compute sentiment scores. 5. **Visualization** Create charts, word clouds, and geographic choropleths to reveal patterns. 6. **Interpretation & Insights** Summarize findings, identify hotspots, and highlight actiona …