SmallholderHub MIS

Context

Programme teams at WRI Indonesia work with oil-palm smallholders, both independent (swadaya) and ex-plasma, and with their farmer groups. Their programme data (farmers, groups, land parcels, training and harvests) needed one reliable home: a consistent structure, quality checks, and reports that partners can trust.

Where

The programme sites are in four regencies of Riau. Regions follow the official BPS hierarchy (province, regency, district, village), and each user’s data access is scoped by province, regency and farmer group.

The challenge

The data comes from the field in different shapes, it holds personal details, and very different people rely on it.

• Mixed inputs. Farmer lists and harvests arrive as Excel sheets; land parcels and oil-palm trees as shapefiles.
• Sensitive data. National ID numbers (NIK) and birth dates must stay masked on screen.
• Many audiences, one source of truth. Five roles, from operators entering records to view-only donors, each seeing only their own regions and groups.
• Fire risk. Hotspots near programme areas need to be flagged and reported quickly.

What I built

I built the platform end-to-end: data model, server, web interface and maps. It is in production.

• Master data with 360° profiles for farmer groups, farmers, land parcels (GeoJSON polygons with revision history) and oil-palm tree points.
• Bulk upload of Excel sheets and zipped shapefiles, with column mapping, a preview and downloadable error rows.
• Dashboards: a main dashboard with 14 summary cards and a clustered map, productivity (t/ha) and plant-age analysis, and training coverage with drill-down to farmers not yet trained.
• Fire alert: NASA FIRMS hotspots classified point-in-polygon against programme boundaries with a 1.5 km buffer, and a print-ready PDF report for all of Riau or per regency.
• Reports and data quality: five reports with Excel and PDF export, and a completeness score for every farmer group.
• Access and security: five roles, a per-menu permission matrix with per-user overrides, soft delete and an audit trail on every table.

The map

Every land parcel is a polygon, so the map explorer is where the programme data comes together.

Outcome

The MIS is live and used by programme teams and partners. Under the hood:

1,200+

commits since March 2026

839

automated tests

22

data models

5

user roles

5

reports (Excel / PDF)

29

in-app help tutorials