Discover and read the best of Twitter Threads about #informationdesign
Most recents (2)
⚠ Warning: this chart is lying to you!
🍬 The chart in this example is trying to suggest a strong correlation between sugar intake and obesity in the US between 1980 and 2000. It does so by carefully choosing the vertical axis ranges so both lines nicely align.
(1/5)
🍬 The chart in this example is trying to suggest a strong correlation between sugar intake and obesity in the US between 1980 and 2000. It does so by carefully choosing the vertical axis ranges so both lines nicely align.
(1/5)
🧐 But with a closer look we can see something else is going on. Sugar intake levels are rising by 30% (from 85g to 110g), while obesity prevalence is rising by 164% (from 14% to 37% of the population). For an accurate comparison, these lines shouldn't nicely align at all!
(2/5)
(2/5)
📊 I've created two redesigns to present some better solutions for this visualization. In redesign 1 we focus on a presentation which is as truthful as possible, comparing the data with the recommended intake level, and enabling an accurate estimation of the prevalence %.
(3/5)
(3/5)
⚠ Warning: this chart is lying to you!
🍬 The chart in this example is trying to suggest a strong correlation between sugar intake and obesity in the US between 1980 and 2000. It does so by carefully choosing the vertical axis ranges so both lines nicely align.
(1/5)
🍬 The chart in this example is trying to suggest a strong correlation between sugar intake and obesity in the US between 1980 and 2000. It does so by carefully choosing the vertical axis ranges so both lines nicely align.
(1/5)
🧐 But with a closer look we can see something else is going on. Sugar intake levels are rising by 30% (from 85g to 110g), while obesity prevalence is rising by 164% (from 14% to 37% of the population). For an accurate comparison, these lines shouldn't nicely align at all!
(2/5)
(2/5)
📊 I've created two redesigns to present some better solutions for this visualization. In redesign 1 we focus on a presentation which is as truthful as possible, comparing the data with the recommended intake level, and enabling an accurate estimation of the prevalence %.
(3/5)
(3/5)
