The goal of ftsAnalysis is to provide a comprehensive toolkit for analyzing humanitarian funding flows using open data from the OCHA Financial Tracking Service (FTS). It offers a set of metrics, indices, analysis, and visualizations to generate insights about funding dynamics.
Language Models are used to automatically generate narrative and reports from three key perspectives:
- Donors: Analyze funding behavior, consistency, and strategic alignment.
- Recipients: Evaluate funding stability, diversification, and dependency.
- Destinations: Assess funding coverage, gaps, and risks for specific crises or locations.
For Developers
This package is built with the help of {fusen} package which allows to maintain consistent documentation through notebooks ( cf dev folder). You can install it from GitHub with:
# install.packages("pak")
pak::pak("Edouard-Legoupil/ftsAnalysis")
# Report generation examplke
ftsAnalysis::generate_report(type = "donor", name = "Switzerland, Government of")To leverage the AI-powered data storytelling features (automated narrative generation for plots), you need to set API keys in your environment (e.g., in your .Renviron file) to configure access to Large Language Models through Azure, OpenAI, Gemini or Anthropic. Alternatively, you can use Ollama for local inference, leveraging Open Source & Reasoning Small Language Models like Gemma3 or DeepSeek-R1.