So I was going to write a blog post, but got distracted by my new baking equipment, including an absurd number of piping tips.

What to do? Data visualization lessons through the medium of brownies + buttercream! A https://abs.twimg.com/emoji/v2/... draggable="false" alt="đŸ§”" title="Thread" aria-label="Emoji: Thread">

(Thanks to @powersoffour for the inspiration)

(1/12)
Side note: I was going to call them visualization tips, but it got too confusing to distinguish visualization tips from piping tips, so I gave up. https://abs.twimg.com/emoji/v2/... draggable="false" alt="😂" title="Gesicht mit FreudentrĂ€nen" aria-label="Emoji: Gesicht mit FreudentrĂ€nen">
Lesson 1: Use colours and different piping tips to distinguish subgroups in your data. Different colours/textures help the viewer separate out groups.

Pictured: practicality (P) versus fun (F) of my new piping tips for star, round and other kinds of tips.

(2/12)
Lesson 2: Avoid chart junk. Putting too many and especially unnecessary elements on your chart makes it hard to read. Also, the ratio of cake to frosting will be off, and nobody wants that.

Pictured: same practicality versus fun of piping tips, but with extra piping.

(3/12)
Lesson 3: Work to make your visualizations more accessible with colour blind-friendly food colour maps, multiple channels (e.g. colour/texture) encoding the same information + alt text.

Pictured: Google trends for the term “cookie” versus “sourdough” for the past year.

(4/12)
Lesson 4: Follow the principle of proportional frosting (apologies to E. Tufte). For area charts, the amount of frosting should scale with the quantity encoded.

Pictured: my projected use of round vs. star piping tips. Encoding is in radius, not area, and hence distorted

(5/12)
Lesson 5: The y-axis on bar charts should include zero. Pretty much an application of the law of proportional frosting.

Pictured: number of round (R), star (S) and other (O) piping tips in my set. Why do I have so many star tips?

Inspired by: @callin_bull& #39;s case study

(6/12)
Lesson 6: Pie charts are for proportional data only. If it doesn’t sum up to 100%, it’s not appropriate for a pie chart.

Pictured: Reasons for buying sweet treats, according to a survey.

Source: https://www.webstaurantstore.com/blog/2122/americas-love-of-baked-goods.html

(7/12)">https://www.webstaurantstore.com/blog/2122...
Lesson 7: Use Gestalt principles like proximity, similarity, connectedness and enclosure to help guide the viewer and make your visualization more readable.

Pictured: just fun brownie bites illustrating Gestalt principles.

inspired by: https://www.nature.com/articles/nmeth1110-863.pdf

(8/12)">https://www.nature.com/articles/...
Lesson 8: Pick the right type of visualization for your type of data.

Pictured: sales of doughnuts (D), cakes (C) and cookies (Co) in 2018. Not the best use of a line chart, because the x-axis is categorical, not numerical.

Source: https://www.webstaurantstore.com/blog/2122/americas-love-of-baked-goods.html

(9/12)">https://www.webstaurantstore.com/blog/2122...
... and then I ran out of both brownies and frosting, or I would have made more.

Here’s a picture of all of them together:

(10/12)
Other lessons I learned: (1) I need more practice with piping, (2) piping numerically accurate plots is *really* hard. (3) Swiss meringue buttercream is not great for piping intricate details, (4) cake-based data visualization is a medium that deserves more attention.

(11/12)
Next week, it& #39;s the first of advent, so I get to start baking Christmas-related treats! Join me for an exciting exploration of three of my favourite things: baking, Christmas and statistical distributions.

The end.

(12/12)
You can follow @Tiana_Athriel.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: