image
image

Land application field trial data

Intro

This dataset is a repository of results from our Land Application Detection Model trial with two organizations. Land application is the process of disposing of agricultural animal waste by spraying it onto fields. We developed a model to detect these practices. This dataset represents the results of a real world trial to verify and label these detected spreads.

Data description

Structured data

  • sent_to_wdnr.csv
    • Each row is a detected spread that we forwarded to our partners at WDNR
  • sent_to_elpc.csv
    • Each row is a detected spread that we forwarded to our partners at ELPC
  • wdnr_responses.csv
    • Each row is a response to a detection from sent_to_wdnr.csv which contains a preliminary determination by WDNR staff as to whether the image looks like a spread and if it was determined to be likely spreading, the results of an investigation into said spread.
  • elpc_responses_raw.csv
    • Each row is a response to a detection from sent_to_elpc.csv which is the results of the ELPC investigation into that detection through the use of citizen volunteers verifiying in person.
  • elpc_responses_clean.csv
    • Same as the raw file but with corrected detection ids to deal with a data entry error.

Image data

  • images/
    • This directory contains .jpeg images of satellite data fed into the model that were sent to either of the partners. Images were captured by Planet using the PlanetScope sensor, visual spectrum 3m images.

Citation

@misc {stanford_regulation,_evaluation,_and_governance_lab_2024, author = { {Stanford Regulation, Evaluation, and Governance Lab} }, title = { land-app-trial (Revision b3d0e11) }, year = 2024, url = { https://huggingface.co/datasets/reglab/land-app-trial }, doi = { 10.57967/hf/1733 }, publisher = { Hugging Face } }

Downloads last month
4
Edit dataset card