Data & BI Analyst · Analytics Engineering · Davao City, PH
Most reports tell you what happened last month. My work answers what to change — where working capital is stuck, which products lose money once discounts are counted, which costs are drifting off budget.
Findings from public-data builds, each documented end to end below.
The pipelines underneath — ingestion, modeling, history tracking — are how I make those answers trustworthy enough to act on. I started as a Credit & Risk Analyst, then ran branch operations analytics: inventory, expenses, liquidations, staff performance. After seven years teaching, I came back to data full time.
How I build
Every platform I ship follows the same spine. Load only what changed, clean it once, and hand the business a model it can query without asking me first.
SELECT * FROM dbo.FactStream WHERE stream_timestamp > '@last_watermark'
Metadata-driven ingestion. One parameterised pipeline loops every table instead of one pipeline per table.
Auto Loader streams bronze into Delta with checkpoints, so a rerun never double-counts. Shared transforms live in one class.
Kimball star schemas with real history. SCD Type 2 dimensions keep yesterday's truth queryable next to today's.
Selected work
Each repo ships with the architecture, the code, and the validation queries I used to prove it works.
Featured case study
Other selected work
glcapitan / fabric-medallion-retail-pipeline
An end-to-end retail analytics platform: ingestion, transformation and dimensional modeling automated, so Power BI reports run off a governed warehouse — and promote between environments without a rebuild.
glcapitan / azureproject
A warehouse that can still answer what a customer's plan was last quarter. Incremental loading keeps it cheap; Type 2 history keeps yesterday's truth queryable — both proven with a live change test.
glcapitan / tableau-sales-customer-dashboard
Found the point where discounting stops paying for itself, and which products quietly destroy margin — across 9,994 transactions, in two linked dashboards driven by one shared parameter.
glcapitan / financial-performance-dashboard
Replaces a manual monthly P&L reporting cycle: change one date cell, refresh, and every KPI, variance and chart re-points against budget, prior period and prior year.
glcapitan / microsoft-fabric-nyc-taxi-analytics
Keeps a high-volume dataset current without a full reload every run — metadata-driven pipelines into a SQL Warehouse, with measures defined once in the semantic model rather than per visual.
glcapitan / fintech-analytics
Rare-event analysis where the signal is a fraction of a percent of the rows: which transaction types carry fraud, and what separates a fraudulent transfer from a normal one.
Stack
Grouped by how deeply I actually use them, not by how impressive the logo looks.
Beyond the tools
Exposure, collections behaviour and portfolio quality — turned into the reporting the business made lending decisions from.
Sitting with the people who own the process and turning "we need a report" into a defined grain, an agreed metric, and someone accountable for it.
Warehouse stock levels and inventory monitoring in branch operations; a lakehouse-to-Power BI build that surfaced ~$95K in trapped capital.
Branch expenses, liquidations and staff performance on daily KPI monitoring — the numbers a supervisor is answerable for every morning.
Pattern analysis across 6.3M PaySim transactions in PostgreSQL and Python — rare-event detection where the signal is a fraction of a percent of the rows.
Case studies and briefing decks written for the people who approve the budget, not for other analysts.
Path
Dashboards, data models, and end-to-end platform builds for clients — plus the public portfolio projects above.
Six months on a live client account: five years of transactional sales data analysed for trend and seasonality, product and category performance, customer segmentation and per-city regional performance — delivered as a Power BI report with findings and recommendations.
Seven years translating complex material for people who did not ask for the technical version — communication, training and stakeholder management. It is now how I run dashboard walkthroughs and document what a metric actually means.
Ran branch analytics end to end: warehouse inventory levels, branch expenses, liquidations, stock monitoring and staff performance — daily KPI monitoring in Excel, VBA and Power BI.
Owned the credit and risk data — exposure, collections behaviour, and the reporting the business made lending decisions from.
Credentials
EXCELHelpline
Microsoft
dbt Labs
Contact
Open to BI and analytics engineering roles, and to freelance work. Based in Davao City, working across time zones. Tell me what you are working with and I will tell you honestly whether I am the right fit.