Not beautiful. More "interesting data set." Source: https://wonder.cdc.gov/ucd-icd10-expanded.html
edited to correct off-by-one error in 5-14 year old column
this post was submitted on 29 Aug 2025
127 points (98.5% liked)
Data is Beautiful
2756 readers
59 users here now
Be respectful
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Huh, these graphs don't visually align in the way i would imagine. Like in the Rate graph at age 84+, heart disease is just slightly higher than cancer, but in the Proportion graph it's MASSIVELY higher.
Can someone give me a simple explanation of why these differ so much?
The bottom data is log so differing scales
OHHHHHHHH! So obvious after you point it out, but so easy to miss on a quick glance. Always look at the axes!
Think of the bottom graph as the absolute, the basis.
Now, think of the top chart as a pie chart of proportional %'s of that basis, for each age group... but its not a bunch of pie charts, its a line plot, where the height of each point = the size of each pie slice.
So if every line is at 10 on the bottom chart, then every line is at 20% on the top chart, because... 5 categories, each is 10, thus each is 20% of the total.
...
The other reason they may seem not to match to you is that the bottom chart is log scale, not linear scale.
It is
0 1 10 100 1000 10000
Not
0 10 20 30 40
OP likely went with log scale for the bottom chart because if you did this linear scale...
It would basically just be a smushed together rainbow of lines at the bottom that then sudden blows out into green and brown as cancer and heart failure start killing everyone in their 50s/60s onward.
(EDIT: yep, they actually did a linear scale version, and its as I said lol)
....
The top chart though, is %'s.
%'s of all total deaths in that age bracket.
It thus... must be percentages, as... displaying %'s ... on a log scale... would be very weird.
Like... you could do it... I guess?
But I've been doing data analysis and making reports and charts and shit, and reading them, for a decade+, and I don't think I have ever seen anyone plot a proportional % on a log scale.
The death rate is on a logarithmic scale