Education Dropout & Budget Insights Dashboard
A dashboard built around a decision rather than a dataset: it joins school-dropout counts, regional budget allocations and university admission records, so a policymaker can see which study programmes to prioritise and whether a district's budget matches its dropout problem.
A Data Lakehouse final project built around a question rather than a dataset: if a district has a dropout problem and a fixed budget, what should it do about it? Three public sources were pulled together to answer that, regency and city budget allocation, school-dropout counts by education level, and university admission records including applicant background. Pentaho Data Integration handled the load from CSV into the database, and the data was modelled dimensionally before any visual was drawn.
The reporting is deliberately decision-shaped. A dashboard that only reports numbers leaves the reader to invent the question, so each page here answers one. The regional page pairs dropout counts against budget released, so the two can be read together rather than separately. The parent profile page carries occupation, education and income distributions, because the interesting question about a dropout is rarely the dropout alone. The programme page shows what applicants actually chose, which programmes were oversubscribed, and how that choice varies by region.
A natural-language Q&A visual sits alongside the built reports, so a reader can ask the model a question directly instead of waiting for someone to build them a page. The people who need this analysis are not the people who build reports.
- Loaded three public Indonesian datasets through a Pentaho Data Integration transformation into a modelled database: 468 rows of regency and city budget allocation, 549 rows of dropout counts spanning primary through vocational level, and university admission records
- Built a three-page report: a regional overview carrying dropout and budget KPIs, a parent socio-economic profile covering occupation, education and income, and a study-programme page with admission outcomes, the top five programmes and their spread
- Added a filled map for programme distribution by region and a natural-language Q&A visual, so a non-technical reader could interrogate the model without touching the report
- Framed every page around a decision: which study programmes the government should prioritise, how to address dropouts within the existing district budget, and where each programme's graduates could be directed