Investigations & Case Studies

All Investigations

Investigating why some Power BI slicers appear to work while others silently fail and how semantic model design determines report behaviour.

An investigation into why candidate-level forecasting requires more than a simple Power BI solution change.

Investigating the trade-offs between Power BI, semantic modelling and application-layer design for delivering analytics experiences at scale.

Users can slice and dice data across multiple attributes. Can we surface a clear summary of the active selections? And if so, how best do we do this?

We evaluated five options, built a prototype using Perspectives and shipped a simpler self-service experience.

Fees can be viewed by created, issued or recognised date. Could we add 'due date' without breaking the existing control flow?

Different employees had targets set for different frequencies (weekly or monthly). Is there anything we could do to persist this frequency selection on drillthrough?

Deleted or archived? Our Analytics logic couldn't tell the difference. These are the questions we had to ask.

A dual-role user needs two different views of the data. A write up on what you can do and what you can't.

How to display a continuous time series line chart without breaking date filter behaviour (including zero values for missing data).

Adding user-controlled weightings and nested drillthrough to a forecast report with trade-offs and implementation notes.

A case study exploring how to add functionality for viewing historic performance

Is it possible to pin Power BI dashboards to Dynamics?

A walkthrough on identifying and fixing slow Power BI reports.

A practical walkthrough of implementing dynamic KPI selection in Power BI.

How to handle Customers having different filtering requirements

Investigation into whether it was possible to build an actuals vs target report using Power Query

A clear five-step workflow that deals with the situation when a customer has multiple production workspaces.

Creating a slicer when the underlying categories are not mutually exclusive and instead overlap is a modelling layer solution not a slicer setting.

Three migration strategies for a core data model change and the new reporting opportunities that are unlocked.

My unsolicited view of the existing data architecture.