Project · precision livestock · Skadia

From farm spreadsheets to a four-metric dashboard

Weights, weather, body condition and treatments usually sit in Excel. This project loads those files and computes four numbers that matter on the ranch: daily gain, animals per hectare, heat-stress hours, and the share of thin cattle.

The dashboard below is already computed. GitHub has Docker to replay it (Airflow + PySpark). Spark runs on a laptop if there is no cluster.

Apache Airflow
PySpark
Delta Lake
Databricks
Docker Compose
Streamlit
Python

GitHub repository Full-screen dashboard

Published result

For operations

ADG — kg gained per day between weighings
Stocking — livestock units per usable hectare
THI — hours of heat stress
BCS — % of animals below body-condition score 4 (scale 1–9)

For data engineering

Bronze / silver / gold layers on Delta. One Airflow DAG, PySpark, and quality checks (keys, nulls, BCS 1–9). Demonstration dataset.

Docker Desktop, about 4 GB RAM. Airflow http://localhost:8088 (admin/admin). Streamlit http://localhost:8501.

git clone https://github.com/FabricioNicolasDuarte/skadia-data-engineering.git

docker compose up -d --build && docker compose run --rm pipeline run-all