Calgary Civic & Housing Analytics Platform
38+ dashboards · property tax, housing, crime, schools, labour
- Designed and built a multi-topic open-data platform for Calgary, covering housing, crime, education,
public health, business survival, and migration.
- Created star-schema models and DAX measure layers to reuse metrics across 38 Power BI pages.
- Implemented monthly and yearly refresh pipelines using Oracle + shell scripts, with documented data
dictionaries and governance notes.
Youth NGO Analytics — Community & Program Insights
Multi-page Power BI · 100k+ synthetic client & volunteer records
- Built an end-to-end analytical solution for a Calgary youth mentoring NGO.
- Designed cohort and engagement dashboards, geographic segmentation, and trauma / risk indicators.
- Engineered a clean data model and reusable DAX measure layer to support multiple story-telling pages
(clients, volunteers, matching, outcomes).
- Produced presentation materials and a 5-minute video focused on insights and real-world program decisions.
Telecom — Yearly Peak Call Analysis (Hadoop / Hive)
Log ETL & performance tuning · historical peak in < 2 minutes
- Originally, finding yearly peak traffic required ~30 person-days of manual Oracle work.
- Built an automated pipeline: Hive DDL and partition loading, and a single query to compute yearly peak.
- Reduced end-to-end processing to ~90 minutes (including transfer + load), with the final peak query
running.
- Designed the solution to be reproducible on a local Hadoop Hive cluster for future benchmarking.
Telecom — Incremental CDR Verification for Customer Call CDR
Partitioned CDR pipeline · ~30× scan reduction · revenue recovery
- Re-designed the monthly customer call CDR check from a full-table scan into a daily incremental process
using Oracle partitions.
- Only scanned the previous day’s partition and marked records as “used” in a status table; end-of-month
work handled only the small remaining set.
- Eliminated query timeouts and reduced OLTP pressure by ~30× compared to the original approach.
- Helped uncover and recover approximately $230,000 in fraudulent commission claims within the first month after deployment.