• Beacon@fedia.io
    7·
    9 days ago

    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?

      • Beacon@fedia.io
        4·
        9 days ago

        OHHHHHHHH! So obvious after you point it out, but so easy to miss on a quick glance. Always look at the axes!

    • sp3ctr4l@lemmy.dbzer0.comEnglish
      10·
      9 days ago

      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.