Behnam Analytics

About

I'm Behnam Ebrahimi, a Senior BI Analyst in the UK. I build reporting that people rely on when the pressure is real: clear metric definitions, models that hold up, and dashboards that lead to a decision.

My work sits between analytics engineering, semantic modelling, KPI design, and decision-focused reporting, mostly in healthcare. I care less about flashy output and more about whether the numbers are defined clearly, trusted by the people who use them, and useful on a bad day.

What I work on

  • Healthcare performance reporting. Waiting lists, patient flow, capacity, and the KPI frameworks behind them.
  • Power BI and Microsoft Fabric. Semantic models, DAX, TMDL, and report design for operational and executive audiences.
  • Analytics engineering. SQL models, data tests, and metric definitions kept as code so dashboards stop breaking quietly.
  • Forecasting. Demand forecasts with honest backtests and prediction intervals that planners can act on.
  • Applied AI. Prompting, agent harnesses such as Claude Code, and review habits that make AI output safe to ship.

How I work

  • Define the metric before building the chart. Most reporting arguments are definition arguments in disguise.
  • Show the caveats. Late data, coding changes, and small numbers go on the page, not in a footnote nobody reads.
  • Test the data, not only the code. Row counts, keys, and accepted values are checked every run.
  • Keep AI on a short leash. It drafts, I review. Anything that reaches a stakeholder has been read by a person.

Tools

Power BI, DAX, TMDL, Microsoft Fabric, SQL, Python (pandas, scikit-learn, statsmodels), Excel, Git, Docker, and Claude Code.

On this site

  • Work: runnable projects on synthetic healthcare data, plus tools I’ve built.
  • Writing: guides on Power BI, forecasting, machine learning, analytics engineering, and AI workflows.
  • AI playbook: prompts and workflows I use for analytics work.
  • Resources: datasets, DAX, TMDL, and code to download.

Healthcare projects here use synthetic data. Nothing on this site comes from a real patient record or an employer’s system.

Contact

Email [email protected] for work enquiries, or find me on LinkedIn and GitHub. Older writing lives in the notebook archive.