Seasonal Spirals(observablehq.com)
observablehq.com
Seasonal Spirals
https://observablehq.com/@yurivish/seasonal-spirals
12 comments
This is really quite nice, but I do believe it would be even nicer if the (IMO ridiculous) insistence on Sunday being the first day of the week was dropped. Right now the visual impact of weekday/weekend patterns is arguably quite reduced, compared to if Monday was the first day.
Per ISO 8601 - the only ISO standard number I can remember - weeks start with Monday: https://en.m.wikipedia.org/wiki/ISO_8601
What does the spiral shape add? I think I personally would find these more readable if the months were straightened out and shown on each their own line. I guess trends wrapping around the cutoff points would be a bit harder to identify, however.
They discuss it in the post, but it provides data density for identifying cyclical patterns. My take is that it’s a fantastic format (alongside more conventional formats, again as discussed in the post) because the way that density is presented is particularly consumable on a small screen.
The most spectacular (and frightening) us is for the climate change:
http://www.climate-lab-book.ac.uk/spirals/
http://www.climate-lab-book.ac.uk/spirals/
This is interesting for me - I think there is a lot of space to develop visualizations that highlight the seasonality of events.
At work, I am often asked to tell stories with data where periodic or seasonal behavior plays a major role. Something like this is definitely improves in many ways over a 7-day moving average or a YoY line chart, but I can't help but feel that there are even more powerful options waiting to be discovered.
At work, I am often asked to tell stories with data where periodic or seasonal behavior plays a major role. Something like this is definitely improves in many ways over a 7-day moving average or a YoY line chart, but I can't help but feel that there are even more powerful options waiting to be discovered.
Seasonality is huge at Amazon. And it depends on the nation too. US, Japan, and EU states all have wildly different seasonalities. Seasonality here should be construed at all levels. For example on a daily basis Japanese peak is after the work day ends, while US has a flatter curve since Americans are willing to shop online during their work day. And then there are special events like the world cup flatlining EU purchasing while US is unaffected.
You might find Cycle Plots [1, 2] useful.
[1]: https://www.cvast.tuwien.ac.at/topics/cycle-plot-climate-dat...
[2]: https://www.perceptualedge.com/articles/guests/intro_to_cycl...
[1]: https://www.cvast.tuwien.ac.at/topics/cycle-plot-climate-dat...
[2]: https://www.perceptualedge.com/articles/guests/intro_to_cycl...
I would argue that while this type of visualization is good for data discovery, it demands too much cognative load for the average audience, at least where I work. YoY line charts or even bar graphs though mundane, require no extra mental effort to the reader and so I find to be more effective in persuasive storytelling.
There are cases in the post’s examples that wouldn’t be achievable with YoY. There’s a lot more information in the spiral presentation, even if it’s fairly fuzzy/low fidelity. They did suggest pairing it with more conventional visualizations, so it’s probably most useful for finding detail that wouldn’t jump out in those viz with common parameters and tweaking the bounds to highlight those.
https://www.youtube.com/watch?v=spUNpyF58BY
The idea is that wrapping a function around a circle at the right frequency (i.e. length per spiral) will cause peaks and troughs to align. It just happens here that the "right" frequency is 1/year.