Ange-Ghislain KOUAMEData Engineer
FR
Delivered

Budget Execution Audit Analytics

An end-to-end audit pipeline on official budget data for a ministry section, from raw spreadsheets to an interactive diagnostic and an automated audit narrative.

Role
Data analyst and BI developer
Period
Mar 2026
Status
Delivered
Context
Ministry of Agriculture and Rural Development · budget section 229, fiscal year 2024 · official State data
  • 363.22 MdFCFA total budget of the section
  • 4administrative units analysed
  • 72.29 %budget consumption rate
  • 100.56 MdFCFA of unspent credits located, with blocking causes diagnosed

Engineering figures describe the system, not its business findings, which remain confidential.

01Executive summary

An analysis of the annual budget execution of a ministry section, built as a reproducible pipeline rather than a one-off spreadsheet: raw Excel data cleaned in Python, documented with the World Bank Metadata Editor, modelled in Power Query and analysed with DAX in Power BI.

The report moves from the overall picture to each administrative unit, then to a root-cause view that separates administrative blockages from financial ones, and generates the audit argument from the measures themselves.

02Pipeline and processing

  • Raw Excel → Python (pandas) cleaning → metadata documentation (Metadata Editor) → Power Query → DAX measures → Power BI Service.
  • Source code and report versioned in a private repository.

03Technology stack

  • Python
  • pandas
  • Metadata Editor
  • Power Query
  • Power BI
  • DAX
  • Power BI Service

04Results and deliverables

  • Three interactive pages: overview, analysis by administrative unit, and a root-cause ('autopsy') view.
  • A consumption rate of 72.29 % on a 363.22 billion FCFA budget, nine critical actions identified, and 100.56 billion FCFA of unspent credits located, with a blockage diagnosed as administrative rather than financial.
  • A performance matrix by structure and critical actions identified for follow-up.
  • An audit narrative generated from the measures, so the text always matches the figures.