MeraolAlemayehu/Inform-africa-Disinformation-Analysis

This repository provides a Python-based framework for analyzing disinformation data. After downloading the CSV from the custom portal, users can upload and analyze it using sentiment analysis, topic modeling, and visualizations. Created by Inform Africa data science team. ## How the Python Code Supports the Framework: User Guide The Python code is designed to assist users in analyzing disinformation data collected through the custom portal. After downloading the data in CSV or JSON format, the Python script can be executed to process, analyze, and visualize the data effectively, supporting the goals of the DISARM framework. Here's how the code works step-by-step: ### 1. **Data Preparation** - **Import the Data**: Start by importing the CSV or JSON file into the Python environment. The code reads the file and converts the data into a format that can be easily processed. - **Preprocessing**: The script performs initial cleaning tasks such as removing duplicates, filling in missing values, and standardizing date formats to ensure that the data is consistent and ready for analysis. ### 2. **Sentiment Analysis** - **Analyzing Post Sentiment**: The Python code uses natural language processing (NLP) libraries to analyze the sentiment of each post in the dataset. It detects whether the post is positive, negative, or neutral, helping to understand the emotional tone of the disinformation. - **Sentiment Visualization**: After the sentiment is determined, the code generates visualizations (like bar graphs or pie charts) to show the distribution of sentiments across all posts. ### 3. **Text Analysis and Topic Modeling** - **Word Frequency**: The code analyzes the most frequently used words or phrases in the posts. By identifying common terms, the code helps uncover recurring themes or topics that could indicate the primary focus of disinformation. - **Topic Modeling**: The script groups similar words and phrases into broader topics, allowing users to quickly identify the main narrative driving the disinformation campaign. This step aids in categorizing the posts by their key themes. ### 4. **Detection of Manipulative Language** - **Identifying Emotional or Manipulative Terms**: The Python code is equipped with algorithms that detect manipul …