Automatic Risk Reporter

This application combines news sentiment analysis and armed conflict data to provide a comprehensive risk assessment platform for different countries and regions.

Features

• News Sentiment Analysis:

  • Search for news articles from the past 24 hours using the Serper API
  • Analyze sentiment of article titles using TextBlob
  • Calculate mean and standard deviation of sentiment scores
  • Summarize the most negative articles using Anthropic's Claude Sonnet 3.7

• Armed Conflict Analysis:

  • Retrieve and analyze data from the ACLED (Armed Conflict Location & Event Data) API
  • Track violent events and fatalities for monitored countries
  • Detect anomalies in conflict patterns using statistical analysis
  • Generate time-series charts of events and fatalities

• Interactive Map Dashboard:

  • Web-based interface with interactive map visualization
  • Color-coded risk indicators for monitored countries
  • Detailed view of news sentiment and conflict data by country
  • Historical data storage in SQLite database

Monitored Countries

The application currently monitors the following countries/regions:

• Ukraine
• Moldova
• Syria
• Lebanon
• Israel
• Libya

Requirements

• Python 3.7+
• API Keys:
  • Serper API (for news search)
  • Anthropic API (for article summarization)
  • ACLED API (for conflict data)

Installation

1. Clone this repository
2. Install the required dependencies:

pip install -r requirements.txt

3. Set up your environment variables in a .env file:

SERPER_API_KEY=your_serper_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key
ACLED_API_KEY=your_acled_api_key
ACLED_EMAIL=your_email_for_acled

Usage

Web Application

To run the web dashboard:

python app.py

The application will be available at http://localhost:5000

Command-line Sentiment Analysis

For standalone news sentiment analysis:

python news_sentiment_analyzer.py "Country Name" --limit 10

Options:

• --limit: Number of negative articles to summarize (default: 10)
• --serper-key: Serper API key (overrides environment variable)
• --anthropic-key: Anthropic API key (overrides environment variable)

Scheduled Analysis

The web application includes a scheduler that automatically runs analysis for all monitored countries once per day. You can also trigger analysis manually through the web interface.

Data Storage

All analysis results are stored in an SQLite database (sentiment_results.db) with the following tables:

• sentiment_results: News sentiment analysis results
• articles: Individual news articles and their sentiment scores
• acled_results: Armed conflict event data and trends
• acled_alerts: Detected anomalies in conflict patterns