Makeover: Sugar consumption

Original visualization


What works with this chart?

  1. The headline is attention-grabbing. Triple the recommendation?!? How is that possible?

What doesn’t work with this chart?

  1. Age breakdown gets a lot of attention (17 groups!). But the message I pull down is that free sugar consumption is well above target for all ages. There are other ways to make it personally relevant.
  2. It’s very hard to get a grasp on the level of consumption - in part due to the percentage targets, and in part because the baseline is daily calorie consumption. Daily calorie consumption is hard because the target varies based on multiple factors including age and gender, and because it is units that are very hard to conceptualize.
  3. For each bar, there’s a “good” amount of free sugar, and excess above and beyond that level. Visually indicating the “safe”, “warning”, and “excessive” levels might bring more urgency.
  4. The headline is about “sugar”, the recommendation and graphic relate to “free sugar” (or in US terms, “added sugar”). That’s potentially a big distinction.
  5. The article is pushing a particular framing – epidemic childhood obesity – and a particular range of policy including tax and plain packaging for soda and candy. That’s a piece of the chart’s import, but a pretty small piece. Might play a guess-game for common products and their added sugar from the nutrition labelling to drive that home.


  1. Data shows the survey detail for two-year survey periods with certain demographic breakouts (children vs. adults, men vs. women, age brackets).
  2. The data seems to have supports for changing trendlines, but is hard to assess. For example, for children or elderly, is there really enough data to make sound conclusions? The total survey is a representative sample of \(N = 1000\) people.
  3. The original data seems to have the actual survey years, rather than this aggregation. Not sure if that really helps, though.

If you torture the data enough, it will confess. - Ronald H. Coase

Potential approaches

  1. Leverage the NHS’s materialization in terms of sugar cubes. (Or the US USDA and HHS recommmendations in terms of teaspoons of added sugar).
  2. Make it personal. Allow configuration based on age and gender to more closely approximate the individual. (Or facet to allow reader to find it.)
  3. Make it interactive. Guess how many calories from natural sugar. Then, free sugar. Surprise with truth!
  4. Choose-your-own-breakfast.

## Response []
##   Date: 2020-01-23 17:54
##   Status: 200
##   Content-Type: application/vnd.openxmlformats-officedocument.spreadsheetml.sheet
##   Size: 9.49 kB
## <ON DISK>  /var/folders/52/6k41kd5123bd8fqvbc9x6nth0000gp/T//RtmpufHg78/file898828321487.xlsx



We might be reading into very little data, but there is at least hope that changes are occurring for children. Enough to go look for more data, perhaps.


Less reason to hope here, and perhaps just seeing the outputs of splitting the dataset so heavily into varying groups.