Methodology & data sources

FedhaFlow is built on official Government of Tanzania budget documents. We process these into a consistent structure so that revenue, recurrent and development figures can be compared across votes, sub-votes and description-level items. This page explains, at a high level, how we source, clean and structure the data, and what limitations users should keep in mind.

Data sources

• Official national budget books (Volume I Revenue, Volume II Recurrent MDAs. Volume III Recurrent RAS, and Volume IV Development).
• Where available, official corrections or updates issued after the initial budget release.

Processing steps

1. Ingestion & parsing – budget books are imported from PDF/Excel, with tables extracted and checked for completeness.
2. Standardisation – column names, codes and formats are standardised so that revenue, recurrent and development data can be analysed side by side.
3. Mapping to levels – each record is mapped to Vote (MDA or RAS), Sub-Vote and Description levels, and tagged by pillar (revenue, recurrent, development).
4. Quality checks – totals are reconciled against official control totals where possible. Basic checks flag missing codes, duplicated lines and obvious formatting errors.
5. Structuring for analysis – final tables are prepared in a flat, analysis-ready format (CSV/XLSX) for use in the dashboard and sector datasets.

Limitations & caveats

• FedhaFlow reflects the approved budget, not in-year expenditure.
• Where official documents contain inconsistencies or errors, we flag them where possible but generally preserve the official figures.
• Some breakdowns are not available at all levels for all years; in those cases, coverage may be partial.
• Users should always read sector- or vote-level findings together with official budget documentation and policy notes for full context.

How to cite FedhaFlow

"Analysis based on FedhaFlow, using official Government of Tanzania budget documents (revenue, recurrent and development)."