Winner of the 2024 Plotnine Plotting Contest

Weather Forecast Plot

Code, Commentary

By John T - GitHub, LinkedIn

We are glad to announce John’s Weather Forecast Plot as the winner of the 2024 Plotnine Plotting Contest.

The judges remarked, “This is a creative use of plotnine to create a visualisation that looks like an actual weather application.” Initially, I was skeptical that it was made with plotnine, and I had to examine the code as it currently shouldn’t be possible to create line segments with a colour gradient. Remarkably, John managed to fake the effect of a gradient on a single segment, and in the submission, we see him exploring his way toward this effect, which is just one element of an exceptionally well-designed visualisation.

The headline plot in this submission is a weather forecast for Madison, WI, whose temperature range falls entirely within the sunny range, making the colours vibrant and the graphic visually striking. Beyond that, this submission simplifies generating forecasts for different locations. Here is a collection of multiple plots captured from the code repository.

As a collection, these plots evoke a range of emotions. There is the satisfaction of having a reusable plotting function; the joy of working out which city we would prefer in those days, is it Madison, Singapore or Tokyo; the chilling angst when we imagine ourselves in Antarctica; then looking at the “testing” plot, we feel the dread of that bug that makes the segments go awry and overshoot their expected ranges, stirring some conditional empathy for the developer.

Visualisations do more than just present graphical information for interpretation; they can take us places. This is a power often overlooked in the realm of statistical graphics, and by crafting this conventional visualisation using unconventional means, we are reminded of that power.

Coal Production Since 1900: Annotated Area Charts with plotnine

Code, Commentary

By Nicola Rennie - website, GitHub, Twitter, Mastodon, Bluesky, LinkedIn

An area plot should not wow us; it is a common type of graph. Yet, this submission of an area plot does just that. The judges remarked, “It is self-explanatory, and moving the y-axis to the right was a brilliant decision, as that is where all the growth is, keeping the information cohesive so that the reader doesn’t have to track across the figure”. They also noted, “It is a beautiful plot with excellent use of colour, annotations, and lines.”

I want to also highlight and appreciate a subtle nuance in this visualisation. Axis ticks are typically short to avoid looking awkward, and horizontal axis labels are preferably centered on the short axis ticks for the same reason. In this case, the ticks are unusually long, but the oddness is mitigated by dashing the lines all so that the year labels can be left-aligned with respect to the lines to prevent any overlap with the previous year’s bounds. The subtlety of the thought, the deliberateness of the action, and the resulting effect are truly sublime.

Dimensionality Reduction, Animated

Code, Commentary

By Jeroen Janssens - Website, GitHub, Twitter LinkedIn,

This was the only animation among the submissions, and the judges praised both the process of creating the video and the visual style. The aesthetics of this animation evoke a sense of nostalgia, reminiscent of the retro visualisations that once supported the research and development of digit classification algorithms – those seemingly low-quality snapshots often found in early Machine Learning books.

Despite having to work around a bug caused by changes in an upstream dependency, this submission excels. It shows how effortlessly a domain-specific snapshot graphic can be transformed into an animation that captures the evolution of an entire system, significantly enhancing communication about such complex processes.

Honorable Mentions

For the Gallery

We asked the judges to note whether each submission they evaluated was interesting enough and unique enough if added to the plotnine gallery. For the submissions below, they said yes.

• Weather Forecast Plot - View the Code. By John T GitHub
• Coal Production Since 1900: Annotated Area Charts with plotnine - View the Code. By Nicola Rennie Website
• Dimensionality Reduction, Animated - View the Code. By Jeroen Janssens Website
• Creating a Plotnine Mondrian - View the Code. By Daniel Tan Website
• Advanced Multi Axis Spider Chart - View the Code . By Zvonimir Boban GitHub
• Clockplot: visualizing sleep data around the clock! - View the Code. By Marco Dalla Vecchia GitHub, Valeria Fonseca Diaz GitHub
• Entering the Golden Age of U.S. Women’s Soccer - View the Code. By Timothy A. Model GitHub
• Plotnine Chronicles: Artistic Heatmaps - Where Pixels Play! - View the Code. By Jaspreet Singh Pabla GitHub
• Plotnine Chronicles: Tetris Tactics - Where Pixels Play! - View the Code. By Jaspreet Singh Pabla GitHub
• 3T$ market cap milestone - View the Code. By Diego Garcia Lozano GitHub
• 🎨 Data Symphony: Vitruvian Data Visualization - View the Code. By Jaspreet Singh Pabla GitHub
• Reproducing Rise of AI over 8 decades plot {plotnine} vs {ggplot2} - View the Code. By Greg Chism GitHub
• Rise of Horror Movies - View the Code. By Deepali Kank GitHub
• Creating a Lollipop Plot with plotnine - View the Code. By Alier Reng GitHub
• Seeing Beyond Statistics: Anscombe’s Quartet and the Power of Graphs - View the Code. By Jeroen Janssens GitHub
• Gas Well Forecast: Decline Curve and Production Rate - View the Code. By Jeff Craggy GitHub