Engineering ·
Blueprint: designing databases for state-scale scheme reporting
An engineering note on the schema patterns behind fast, consistent reporting across districts, schemes and years.
Scheme reporting combines high write volume during enrolment with heavy read load during reviews. Designs that handle one well often struggle with the other.
This note covers the patterns Sumato uses: a normalized operational core, master data for beneficiaries and geography, partitioning by time and district, and a separate analytical layer for dashboards.
It also explains how duplicate beneficiaries are detected across schemes without slowing down enrolment.