Skip to main content
Zenodo

Farm-Flow — AG-IoT Intrusion Detection Dataset [1.31M Flows, Smart Agriculture]

Smart Agriculture
1 views
2 min read
License

Abstract

"Agricultural IoT network intrusion dataset with 1.31 million labeled flow records (532 MB) emulating a real AG-IoT farm environment. Covers crop health, weather, and soil condition data with network attack scenarios. CSV via Zenodo. Used for smart farming security research."

Description

Overview

Farm-Flow is a purpose-built network intrusion detection dataset designed to emulate real-world Agricultural IoT (AG-IoT) systems. It was created to address a critical gap in IoT security research: despite rapid adoption of precision agriculture technologies, no publicly available intrusion detection dataset existed that specifically modeled the threat landscape of smart farming environments.

The dataset simulates an AG-IoT infrastructure incorporating soil condition monitoring, crop health tracking, weather stations, and irrigation control sensors — the typical deployment topology of a connected farm. Network attack traffic was generated across this simulated farm environment and processed through comprehensive cleaning and feature extraction pipelines.

The resulting dataset contains 1,310,000 labeled flow instances totaling 532 MB, and achieves an intrusion detection accuracy of 92.67% under benchmark testing. Farm-Flow is hosted on Zenodo under an open license and is available for direct download, making it freely accessible for smart agriculture cybersecurity research.

Column Schema

ColumnDescription
flow_idUnique identifier for each network flow.
src_ip / dst_ipSource and destination IP addresses.
src_port / dst_portSource and destination port numbers.
protocolTransport protocol of the flow.
flow_durationDuration of the flow in microseconds.
fwd_packets / bwd_packetsForward and backward packet counts.
flow_bytes_per_secFlow throughput in bytes per second.
soil_sensor_flagFlag indicating flow originated from a soil IoT sensor.
weather_sensor_flagFlag indicating flow from a weather monitoring endpoint.
labelBinary label: Benign or Attack.
attack_typeSpecific attack sub-category label.

Key Statistics

  • Total Records: 1,310,000 labeled flow instances
  • File Size: 532 MB (compressed)
  • IoT Environment: Simulated AG-IoT farm with soil, crop, weather, and irrigation sensors
  • Benchmark IDS Accuracy: 92.67%
  • File Format: CSV
  • Published: April 2024 on Zenodo

Use Cases

  • Intrusion detection system development for precision agriculture IoT deployments
  • Network attack classification in agricultural sensor network environments
  • Smart farming cybersecurity research and benchmarking ML/DL classifiers
  • Crop health and irrigation control system threat modeling

Source & Attribution

The Farm-Flow dataset was published on Zenodo in April 2024. It is openly accessible for download and was created to support the development of security solutions specifically tailored to Agricultural IoT ecosystems, a growing area of critical infrastructure protection research.

View Data Structure

To explore column names, data types, and sample rows, visit the official dataset page on Zenodo.

Preview on Zenodo

Cite This Dataset

Narjes Davari, Bruno Veloso (2021). Farm-Flow — AG-IoT Intrusion Detection Dataset [1.31M Flows, Smart Agriculture]. [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5VW3R

Source: UCI Machine Learning Repository (2021) · DOI: 10.24432/C5VW3R

Indexed by IoTDataset.com on Apr 17, 2026

Ready to Start Your Research?

Download this dataset directly from the official repository and start building your next breakthrough project.

Download Dataset

Related Topics & Keywords

Share This Research

More in Smart Agriculture

View All