← All work
LiveMasinyusane · Zazi iZandi
Data portals with an AI analyst built in
Public and internal dashboards over four years of assessment, session and youth-programme data, with chat agents that answer questions about the results.
- children in the data platform
- 10,700+
- children in the data platform
- schools
- 341
- schools
- session records
- 134,000+
- session records
The problem
Programme data lived across Airtable bases, spreadsheets and field tools. Leaders and funders needed trustworthy numbers, and managers needed answers without waiting for an analyst.
What I built
- A central data warehouse that syncs nightly from the field systems, with stable IDs for every child and school across years.
- Public and staff-only data portals with dashboards for literacy assessments, sessions and youth programmes from 2023 onwards.
The AI inside
- MasiAnalyze: a chat analyst that uses function calling to query real programme metrics and grade-level benchmarks, instead of guessing.
- Zazi Bot: an agent built on the OpenAI Agents SDK that answers questions about results for logged-in staff.
Where it stands
Both portals are live (data.masinyusane.org and the Zazi iZandi portal), and the AI analysts are available to signed-in staff.
PythonStreamlitDjango + PostgresOpenAI function callingOpenAI Agents SDKDocker
Want something like this?
Tell me what you're trying to do. If I can build it well, I'll show you how; if not, I'll point you somewhere better.