Spatial Sales Intelligence

Problem

PT. Makassar Raya Motor runs 10 Daihatsu branches across South, Central and Southeast Sulawesi. Sales decisions were made without a clear geographic view of where demand was strong, weak, or untapped.

Context

Sales data existed, but not tied to geography or to socio-economic context. The analysis had to work at multiple levels — province, district, sub-district and village — and be usable by non-technical stakeholders.

Solution

• Spatial analysis of sales across 3 provinces, 54 districts, 709 sub-districts and 7,394 villages.
• Maps with context — combined with GDP, per-capita income and population indicators from BPS, designed together with urban-planning experts.
• Interactive dashboards built with Python Plotly.
• A full-stack web app (Laravel Livewire) for data management, imports and the embedded dashboards, backed by a FastAPI service for spatial processing.

Architecture

Technical challenges

• Absorbing and cleaning GeoJSON for three provinces at several administrative levels.
• Keeping maps responsive while rendering hundreds of sub-district polygons.
• Joining sales records to administrative areas consistently.

Impact

provinces analysed

3provinces analysed

sub-districts mapped

709sub-districts mapped

villages covered

7,394villages covered

Stakeholders could explore sales performance next to socio-economic indicators at a glance, instead of reading spreadsheets per branch.

Lessons learned

The value of a dashboard is in the questions it lets people ask. Pairing sales with public socio-economic data turned "where did we sell?" into "where should we sell?".

Snapshots

Map analysisDashboard analysisData managementData import feature