May 2026 – Aug 2026
Capital Projects Performance Dashboard
A live 10-page Power BI dashboard that gives leadership one view of 100+ capital projects.
- Power BI
- DAX
- Power Query (M)
- Python
- Star schema
- Data quality checks
Illustrative pipeline
Illustrative pipeline. Generic teaching example, not internal Novolex architecture.
- Data Source: A live feed of project status arrives on a schedule.
- Power Query: Power Query and Python prepare the data for the model.
- Data Model: A generic star schema connects measures with their context.
- DAX: Measures turn the model into totals and rates.
- Dashboard: Pages present the results alongside data quality checks.
- 10 pages
- 100+ projects
- 250+ milestones
- 25+ DAX measures
- 8 automated checks
Problem
Leadership needed one live view of a growing portfolio.
With 100+ capital projects and 250+ milestones, leadership wanted one reliable place to see which projects were on track, how milestones were trending and where to focus attention.
Role & Action
I started with the decisions, not the data.
I met the business team to learn which decisions the dashboard had to support and which questions it had to answer, then built it in short feedback cycles, refining the design each round.
How the feedback rounds worked
Each round I showed progress, walked through what I had built and why, listened to how they would actually use it, and refined the design before the next round. The result fits how the team already runs its reviews.
Built with the people who use it
Ask which decisions the dashboard must support.
Workflow
A pipeline and model built to stay reliable.
A Power Query (M) and Python pipeline on a live, auto-refreshing source feed flows into a star schema data model, which keeps filters and time-based measures reliable.
Inside the pipeline
I built the pipeline with Power Query and Python to support the dashboard's regular refreshes.
The measures
I wrote 25+ DAX measures for the dashboard's reporting needs.
How data becomes a dashboard
Simulation with made-up data. Illustrative simulation with fictional data. Not Novolex data or internal rules.
- Watch the pipeline build automatically.
- Hover a dimension to follow a filter.
- Move the slider to change when a milestone counts.
Illustrative pipeline
Illustrative pipeline. Generic teaching example, not internal Novolex architecture.
- Data Source: A live feed of project status arrives on a schedule.
- Power Query: Power Query and Python prepare the data for the model.
- Data Model: A generic star schema connects measures with their context.
- DAX: Measures turn the model into totals and rates.
- Dashboard: Pages present the results alongside data quality checks.
Star schema, generic example
Hover or focus a dimension to follow its filter.
Role & Action
Automated checks keep the numbers trustworthy as it grows.
I designed 8 automated data quality checks, so reporting stays accurate as the portfolio scales and new data arrives.
Built to last
I made it future-proof so it keeps working as new data arrives, and documented the full pipeline and data model in a formal spec so the team can maintain and extend it.
When does a milestone count as met?
Illustrative simulation with fictional data. Not Novolex data or internal rules.
3 of 6 fictional projects count as met.
Result
Now part of the team's regular project reviews.
Leadership has one live view of 100+ capital projects and 250+ milestones across 10 pages.
Quality built in
Data quality monitoring is part of the dashboard itself, so reporting stays accurate as the portfolio grows.
- Projected reporting work saved per month
- 4 to 8 hrs
- Estimated improvement in status reporting efficiency
- ~25%
I later rejoined Novolex on the IT Emerging Technologies team to build AI agents and Power Platform automations. View the role