A profit and loss statement that renders in 206 ms. The same 28 x 15 grid of numbers took 12,412 ms as a native matrix driven by a calculation group, so I rebuilt it as a single Deneb visual over one query.

The blog article is the backstory: why the matrix was slow, the four approaches I benchmarked side by side, and the DAX and spec behind the version that won.

This report actually helped me to land my first PBI Dev contract with a private healthcare provider.

This is why it is good to have a report showcase portfolio. Even better if your data is related to the industry in which you would like to work.

This report gives insight into customer sales metrics like moving averages, new and returning customers, and satisfaction metrics.

The middle line chart uses a secondary y-axis that consistently shows all sales when dimensions are filtered, so the user can see what proportion of the whole the selected dimensions make up.

This project explores the capabilities of field parameters.

Field parameters are a winning PBI feature for building highly interactive visuals. In this example users can dynamically control the displayed metrics, adjust date granularity, and see appropriate number formatting with the click of a button.

This report focuses on showing variances to a selected metric (target, last year or two years ago).

The KPI in the top right shows an overview of the selected year and all of the comparatives while the line chart gives the detail on a monthly basis. This report also makes use of the very awesome PBI feature, field parameters.

This report makes use of a Top N selection as per business line (department).

This report also makes some great use of field parameters for selecting visible chart metrics. Report users are also able to sort the charts by revenue or gross margin via a slicer.

146 years of global temperature, 1880 to 2025, drawn as one point per year with a smoothed trend running through them. The smoothing follows the data locally rather than forcing a single straight line through the whole record, so what you see is the shape of the warming and not just its average direction. Core Power BI visuals only: the fit is written in DAX, and a slicer widens or narrows the smoothing window.

That shape is the whole argument. The dashed straight line reads the record as one steady climb and calls 2025 at +0.68 °C when the year actually came in at +1.19. The curve lands on +1.10, and it keeps the two features the straight line erases: the pause from 1940 to 1975, and the sharp acceleration that starts in the late 1970s and is still running. Warming has not been constant, and a straight line is the one shape that cannot say so.

A horizontal dumbbell chart comparing average salary against median home price for 15 US states, with the dollar gap labelled above each connector and the states sorted by that gap. Ohio sits at the top on a $200K gap, California at the bottom on $705K.

The whole chart is one Deneb (Vega) visual: subtitle, centred legend, axes, connectors, gap labels, tooltips and source note are all marks in a single spec. The gap is worked out in a Vega formula transform rather than a calculated column, and the same field drives the sort, so the model stayed at six plain columns.

Thirteen regions of the British Isles drawn as one Euler diagram, from the root set down to England, Scotland and Wales. Northern Ireland sits in the overlap, because it belongs to both the United Kingdom and the island of Ireland.

The page is a single Deneb (Vega) visual over a 13 row table of name, level and parent. The levels slicer is drawn inside the same spec rather than as a separate Power BI slicer: clicking a row sets a signal that filters every ellipse and label to a level at or below the selection, and stroke weight thickens toward the root, with yellow marking Great Britain.

Seventeen Superstore product sub-categories, each shown as the difference between its total sales and the sales of one chosen baseline sub-category. Clicking any row in the list on the left resets the baseline and every bar recalculates against it, with a “No Comparison” option that drops back to raw totals.

The whole chart is one Vega specification in Deneb. The baseline is held in a Vega signal rather than a Power BI slicer, so the visual receives only Sub-Category and Total Sales and derives the variance, the descending sort, the labels and the colour itself. Each side runs through a ramp of 21 stops, blue above the baseline and red below, brightening with distance from it.

Forty films across ten four-film franchises, drawn as peas in a pod. Each pod is a franchise, each pea is one film’s IMDB rating, and the pods are ordered by total franchise box office, giving Avengers at the top and Evil Dead at the bottom.

The chart is a Deneb (Vega-Lite) spec in three layers: a pod casing rim, the casing interior, and the peas coloured by release order. Power Query unpivots then re-pivots the film rows into one row per franchise, and a hidden report page tooltip uses REMOVEFILTERS measures so hovering a single pea returns all four films of that franchise. Tied ratings are nudged apart vertically so no pea hides behind another.