An automated, end-to-end data pipeline for processing field survey data, assessing EUDR (European Union Deforestation Regulation) traceability, and generating sustainability KPIs. The project combines an R-based analytics workflow with an interactive Streamlit dashboard for stakeholder reporting.
Coffee sustainability data can be fragmented across different sources and formats, making it difficult to assess traceability, identify data gaps, and monitor sustainability indicators. Incomplete geolocation and inconsistent field records can also make EUDR-related due diligence more challenging.
I developed a modular and reproducible analytics pipeline that transforms raw field data into structured sustainability insights:
- Automated Audit Pipeline: Processes field records and flags entries requiring further action based on polygon-mapping status.
- KPI Derivation: Uses an automated R pipeline to calculate regional traceability rates and sustainability metrics.
- Stakeholder Dashboard: Provides an interactive Streamlit interface for filtering results by region and monitoring key sustainability indicators.
This project is informed by EUDR traceability requirements and common sustainability data-management practices:
- EUDR Traceability: Uses polygon-based geolocation to assess whether farm records contain the spatial information needed for traceability and due diligence.
-
Carbon Modeling: Calculates carbon-intensity indicators using CO₂-equivalent metrics such as
$tCO_2/ha$ . - Data Quality: Applies a data-audit-first workflow that flags quality issues before analysis and KPI generation.
Based on the simulated pipeline results, the following interventions are suggested:
- Prioritize Geo-Mapping: Prioritize field data collection in Gedeo and Sidamo, where the simulated results show lower traceability rates.
- Targeted Training: Consider prioritizing carbon-reduction and sustainable farming training in Sidamo, where the simulated data show higher carbon intensity.
- Data Generation: Python (
Faker,Pandas) for synthetic supply-chain data generation. - Data Processing: R (
tidyverse,janitor) for data cleaning, auditing, and KPI generation. - Pipeline Orchestration: Modular R scripts with error handling using
tryCatch. - Dashboard: Streamlit for interactive stakeholder visualization.
Aklilu Abera | Data Analyst