How to make a bar chart race

Race tools expect one of two layouts. The wide one, used by Flourish and by default here, has a row per competitor: a name, optionally a category (region, league, company group), then one column per period. The long one has a row per value: period, name, value. Spreadsheets of downloads from the World Bank, Our World in Data or a sales system are usually one or the other.

• Use cumulative values if the race should show running totals (total goals so far, not goals that week).
• Keep periods evenly spaced; a missing year inside a series is bridged, and a gap at the start or end makes the bar fade in or out.
• Use the same name spelling in every row of long data, or the tool sees two competitors.

Flourish has a dedicated bar chart race template. Sign in, create a visualisation from the template, and replace the data on its Data tab: names in rows, optional image and category columns, then one column per period. Captions for particular periods go on a second sheet.

Flourish draws an excellent race, but two limits matter if you want a video: free projects are published publicly with Flourish attribution, and the free plan can’t download images or video. MP4 export is a story feature on the Enterprise plan that Flourish switches on for you; otherwise people screen-record the race. If you already have a Flourish race sheet, paste it into the tool above: it reads the same layout.

The open-source bar_chart_race package turns a pandas DataFrame (periods as rows, competitors as columns) into an MP4 or GIF with one function call, using matplotlib for drawing and ffmpeg for encoding.

1. Install Python, then pip install bar_chart_race (and ffmpeg for MP4 output).
2. Load your sheet with pandas and pivot it so each row is a period and each column a competitor.
3. Call bar_chart_race with the DataFrame, a filename and options such as n_bars and period_length.

It’s free and scriptable, which suits regular reports, but fonts, colours and label placement take matplotlib knowledge, and rendering long races can take minutes.

Use Flourish when

the race lives on a web page, people should be able to pause it, and your organisation already pays for Flourish.

Use Python when

you rebuild the same race every week from a data pipeline.

Use the browser tool when

you need an MP4, GIF or vertical video now, without an account, and the data shouldn’t leave your computer. It also exports line chart races and bubble races from the same data.