Demo data sets for Federated Learning forest monitoring app example

This repository contains dummy datasets designed for testing and demonstration purposes in the Federated Learning Forest Monitoring App, hosted on the Flower Hub (flower.ai). The Federated Learning Forest Monitoring App presents an basic example following the preprint: https://zenodo.org/records/17415920

The data allow users to run end‑to‑end simulations of federated training, communication rounds, client–server interactions, and model evaluation - without exposing any real or sensitive information. 

Contents

The dataset includes three NumPy .npz archives:

1. demo_data.npz

   • Contains simulation data intended for the flower simulation mode.
   • Variables mimic the structure and numerical ranges of typical forest monitoring features (e.g., satellite-derived indicators, environmental covariates, or time‑series metrics).
   • All values are artificially generated and do not correspond to any real-world observations.

2. client_1_demo_data.npz

   • Dummy local dataset for Client 1 in a federated learning setup.
   • Structured to resemble a realistic client-side partition similar to the simulation data set above.
   • Enables testing of client computations, local training, and communication routines.

3. client_2_demo_data.npz

   • Dummy local dataset for Client 2.
   • Same structure and purpose as Client 1.

Purpose

These datasets are intended for demonstrating the functionality of the Federated Learning Forest Monitoring App.

Characteristics

• Fully synthetic; no real forest or satellite data is included
• Structured to match the inputs needed for the Forest Monitoring App example on the FlowerHub
• Lightweight and suitable for rapid testing