Bringing years of program data together into a single source of truth for Australia's teacher training program

Teach for Australia (TFA) is a national not-for-profit that recruits, trains and supports outstanding graduates and career changers to teach in under-resourced schools across Australia. Every year, TFA tracks hundreds of associates from the moment they begin their teacher training through to their careers beyond the program.

That tracking relied on a well-maintained set of Excel-based processes. Over time, as the volume and complexity of TFA's data grew, those processes became harder to scale. Bringing together data about associates, schools and alumni across multiple sources required increasing manual effort, and for an organisation that relies on its data to report to government on grant eligibility and track long-term program impact, that effort was becoming a constraint.

TFA needed a single, reliable place to store and report on this data. One that could handle the complexity of tracking people and institutions over time, and surface insights about program outcomes without requiring a data analyst to stitch spreadsheets together every reporting cycle.

Kiandra designed and built a data warehouse for Teach for Australia using Microsoft Fabric.

The solution processes three core datasets: cohort information about associates in each intake year, alumni survey data tracking where graduates are teaching and what roles they hold, and school information used to assess grant eligibility. TFA uploads these as Excel files once a year. From there, the data moves automatically through three processing layers:

• A Raw layer that ingests the uploaded files without modification, preserving the original data with a version-aware structure to handle future template changes
• An Enriched layer that cleanses, validates and applies business rules, producing a structured, auditable source of truth with a full change history
• A Curated layer that transforms the enriched data into a dimensional warehouse optimised for Power BI reporting.

The architecture uses a variation of the Fabric Medallion Architecture across nine separate workspaces, separating development, test and production environments as well as the three processing layers. This structure enforces clear security boundaries: most TFA staff access only the reporting layer, while administrators control data ingestion and pipeline execution.

The solution was built entirely within Microsoft Fabric, a relatively new unified platform from Microsoft that consolidates previously separate tools into a single offering. Kiandra invested upfront in learning and development on the platform to deliver the project, and passed that capability back to TFA through training materials and system walkthroughs.

The technology stack includes:

• Microsoft Fabric (Lakehouses, Data Warehouse, Notebooks, Pipelines)
• Power BI for end-user reporting
• Azure DevOps for source control
• Azure role-based security for access management.

The project was delivered on time and within the fixed-price budget.

Teach for Australia now has a single, maintained data warehouse that brings together all program data into a structured, reportable system. Reporting that previously required significant manual effort can now be run directly from Power BI.

The warehouse supports a range of strategic reports including cumulative teaching placements by cohort, alumni working in TFA-eligible schools, associates teaching STEM subjects, ATSI associate participation, and associate career pathway tracking. This includes the “TFA by the numbers” report used for government grant reporting, which previously required manual effort to produce.

The change history capability within the Enriched layer makes it significantly easier for TFA to track how associate and school data has changed over time - something that previously required considerable manual effort to piece together.

The solution was handed over to TFA with full client documentation, training videos and a two-year warranty arrangement. Kiandra also identified a clear path for future phases, including connecting directly to Salesforce for live CRM data rather than relying on manual exports, and expanding the reporting capability.