Technical analysis of news sequence
-
Updated
Sep 14, 2021 - HTML
Technical analysis of news sequence
This project helps identify whether a news article is Real or Fake using AI and text analysis techniques. Users can enter news content and receive a prediction along with a confidence score.
Multi-source news bias analyzer using NLP, FastAPI, and Chart.js
SAT News Analyser is a web application offering in-depth news article analysis with features like sentiment analysis, genre detection, POS tagging, and reading time estimation. Supports URL-based content extraction, Google OAuth authentication, article history, and text-to-speech functionality. Built with Flask, PostgreSQL, and NLTK.
LLM-vs-digital-trace sentiment analysis project comparing electric vehicle public opinion across Reddit, news, and model-generated sentiment; selected for AAPOR 2025.
Live AI news discussion in the browser. Light and Dark debate current headlines from opposing perspectives, with inline source cards, local WebGPU inference, typing indicators, and continuous news polling.
Real-time geospatial visualization of global news coverage. Tracks country mentions across 100+ headlines, applies intelligent alias matching, and renders an interactive Plotly heatmap. Built with Python, NewsAPI, and Pandas.
Can open tools build a reliable news corpus? Reproducible study of Media Cloud discovery, polite crawling, Trafilatura extraction, and LLM-judged completeness across 30 US outlets.
AI-powered news opinion analysis and manipulation detection system (Naver/Daum News)
An AI-powered application for detecting fake news using machine learning. Features a FastAPI backend with URL scraping capabilities and two frontend options: a modern web interface and an interactive Streamlit dashboard.
Око Леванта — витрина: разборы Ближнего Востока по ивритским, арабским и английским источникам, со списком источников под каждым
To associate your repository with the news-analysis topic, visit your repo's landing page and select "manage topics."