Statewide Property Intelligence for 638,184 Parcels
Venkatesh Shivandi
The challenge
A state-chartered land bank needed to decide which neglected properties to acquire. Ownership, environmental, census, and parcel data lived in separate systems, making decisions dependent on manual research.
What I built
I combined 5 disconnected sources into a single property record and scored every parcel against 45 measures, including opportunity, ownership complexity, and feasibility of purchase. I also created interactive maps for board-level review.
Result
638,184 parcels scored
249,034 ownership records reconciled
205 publicly owned, high-impact parcels identified from the top tier
Constraints
The model ranks opportunity alongside confidence. Census and environmental inputs lag real conditions, and ownership records covered 39% of parcels.
Like this project
Posted Sep 2, 2026
Unified 5 data sources, scored 638,184 parcels, and delivered an actionable acquisition shortlist for a public-sector land bank.
