polldatavisualizer

[polldatavisualizer]

tumblr poll results are often misleading for polls that involve more than one variable

my mission is to create representations of data collected in tumblr polls that allow the data to be interpreted meaningfully, and to improve statistical literacy on this website.

if you’re like me, you’ve probably looked at polls interrogating more than one variable, you know, ones like this:

Q: ARE YOU WEARING PANTS RIGHT NOW?

OPTION 1: Yes, and I’m a horse that uses tumblr - 0.1%

OPTION 2: No, and I’m a horse that uses tumblr - 0.7%

OPTION 3: Yes, and I’m a human - 69%

OPTION 4: No, and I’m a human - 30.2%

and thought to yourself, “man, a bar chart without any kind of normalization is a super confusing and possibly misleading way to represent these data.” or maybe that’s just me.

I feel strongly about good data viz, and I started doing this for myself because I was curious how certain poll results would look if the data were represented better. this is (mostly) a lighthearted blog.

if I reblog your poll and you want my reblog deleted, tell me via the ask box and I’ll delete it—but you need to provide a reason (“this poll was a joke” is a perfectly valid reason, as is “I don’t like your vibe”—you just have to tell me something). this is because lack of statistical literacy and bad data representation being seen by large swaths of young impressionable tumblr users genuinely concerns me a little bit. please note (before you, poll creator, get offended) that tumblr’s polling system is at fault here, not you, poll creator, trying to collect information with a very limited data collection tool.

you can submit polls you want interpretation for and I’ll do my best. rules for submitting polls:

the poll doesn’t have to have closed, but I won’t interpret the data until the poll closes. for obvious reasons.

you can submit joke polls, but I will treat them the same as any other poll.

some polls are formulated in such a way that makes the data collected inherently uninterpretable. if you submit a poll that I feel is uninterpretable, I will say so. don’t get mad, it’s my opinion.

please DO NOT submit polls from fandom-oriented poll blogs (character tournaments, the like). I just don’t care and don’t want to get caught up in drama surrounding those sorts of things.

about the mod: I’m an adult. mid 20s. my pronouns are he/him. I am white, transgender, autistic, and queer. I’m a biologist, which is a field with a serious data misinterpretation and misrepresentation problem, which is why I’ve got this big ass chip on my shoulder.

I do data viz in R using ggplot2, which is overkill for this application, but it’s fun for me, and stupid easy to crank out plots. I’ll publish all code I write for this blog. Learn R! She’s open source! She’s kinda wacky and illogical sometimes as a result but we love her!

no terf shit. no racist shit. no cop shit. no zionists. kill yourselves and go to hell. will do my best to check OPs for these things but if I reblog a poll made by someone who falls into any of these categories please let me know ASAP so I can delete and block.