JoyceAbache/maji-ndogo-part2-powerbi-datamodel

Part 2 # Maji Ndogo Water Services — Part 2: Moulding Data into Visual Stories Second project in a multi-part Power BI series analyzing water access, infrastructure, and safety in the fictional country of Maji Ndogo. This project moves from a single-table analysis (Part 1) to a full relational data model built from multiple linked tables, then uses that model to explore national water access, queue times, well pollution, and water-related crime. ## Project Overview **Goal:** Build a proper Power BI data model from seven related tables, then design an interactive report that lets a decision-maker filter national results down to a specific province and see the full picture — water access, wait times, pollution status, and safety risk — all update together. **Data source:** `Md water services data.xlsx` — seven tables: `visits`, `location`, `water_source`, `well_pollution`, `queue_composition`, `project_progress`, `water_source_related_crime`. ## Data Modeling The tables didn't connect cleanly out of the box, and fixing that was most of the work before any visual could be built: - Corrected column headers that failed to import properly on `location`, which broke its relationships to the rest of the model - Removed unneeded identifier columns from `queue_composition` and `water_source_related_crime` - Resolved a many-to-many relationship between `visits` and `water_source_related_crime` by linking through a table with a unique set of `location_id`s instead - Set relationship directionality deliberately: one-directional by default (to avoid ambiguous filtering), switched to bi-directional only where a specific visual required it — e.g. the `visits`–`water_source` relationship, needed so filtering by province correctly filters water source type - Extracted `day_name` and `hour_of_day` from timestamp columns using Power Query's Date/Time tools, since the raw timestamp didn't support day-of-week and hour-level analysis directly **Result:** a multi-star schema — `visits` is …