Nelvinebi/election-sentiment-tracker

An automated X data pipeline tracking Nigeria's 2026 presidential election sentiment for Tinubu, Obi, and Atiku. Uses NLP to analyse daily mentions, compute 7-day and overall trends, and auto-publish charts with structured write-ups. # 🗳️ Presidential Election Sentiment Tracker: Nigeria 2026 > An **automated X (Twitter) data pipeline** that collects, analyses, and visualises daily sentiment for Nigeria's 2026 presidential candidates **Tinubu (APC)**, **Peter Obi (NDC)**, and **Atiku (ADC)** using real-time social media mentions, NLP-powered sentiment analysis, and publication-ready charts auto-posted to X. --- ## 📌 Problem Nigeria's 2026 presidential election is generating intense discourse across social media, with millions of posts daily reflecting shifting public opinion, campaign momentum, and voter sentiment. Traditional polling is expensive, slow, and often fails to capture real-time grassroots sentiment particularly among Nigeria's digitally active youth demographic. Without automated sentiment tracking, analysts and campaigns face: - **Missed momentum shifts** failing to detect when a candidate is gaining or losing support - **Noisy volume metrics** counting mentions without knowing if they are positive or negative - **Delayed response times** waiting days or weeks for poll results while sentiment moves daily - **Unstructured data overload** drowning in raw tweets without statistical summarisation There is a critical need for a **reproducible, automated, open-source system** that transforms raw social noise into structured electoral intelligence collecting candidate mentions from X in real time, analysing sentiment using Natural Language Processing, computing daily, weekly, and overall statistical trends, and auto-publishing visual summaries to X. --- ## 🎯 Objective - **Ingest** daily mentions of Tinubu, Peter Obi, and Atiku from X API v2 using Tweepy - **Analyse sentiment** of each mention using a pre-trained RoBERTa NLP model (`cardiffnlp/twitter-roberta-base-sentiment`) - **Compute statistical summaries:** daily metrics, 7-day rolling averages, and cumulative trends since day one - **Classify sentiment** into positive, negative, and neutral with compound s …