this post was submitted on 13 Mar 2024
1090 points (93.9% liked)
Technology
59719 readers
2776 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
That graph is trash. The baseline needs to be at zero.
That graph hurts my data scientist heart
It’s common practice to cut the y axis, did you guys not cover that in visualisation?
Could you please clarify why the baseline needs to be at 0? I'm genuinely curious.
This graph gives the impression that the total installation number has been multipliés x4 or X5 while it is not the case when looking at the raw numbers.
Any variation can look impressive if you zoom enough, that's why you need a baseline at 0. This way you see thé entire scale of the phenomenon
How so? It goes from ~7 to ~11. That's not even x2.
Yes but the graph goes from 2 rectangles above the bottom line to 8 rectangles above the bottom line in that final surge.
So visually, it looks like it has quadrupled.
While I agree for the sake of clarity, a bigger problem is that it only goes back less than 2 months. Has the number of installs been steady at 7k for a long time? Or does it fluctuate wildly like this occasionally for reasons totally unrelated to laws?
I was just clarifying the original comment about the baseline not being 0.
Tbh, I hadn't even looked at it properly and only noticed now that the timeline isn't one month per box.
No it doesn't.
It's meant to illustrate a change and it does so perfectly fine. It's not a scientific paper.
It's a 32-34% increase looking at the graph. That's significant enough to shout about.
Imagine any change you could make surprising competition by 25% in any market. That's huge.
Define "perfectly fine". It is clearly exaggerating the change. At a glance it looks more like a 5 times increase, not a 30% increase.
Of lies, damned lies, and statistics this graph is certainly one of them.
Did you know that disco record sales were up 400% for the year ending 1976, if these trends continue...AY!
But you don't get that percentage from looking at the graph. You get that from looking at the numbers.
The graph height increases by 300% in the last ~~3 months~~ 9 days.
You could say the same about a 0.001 difference if you zoom in on the y-axis. You don't know what you're talking about.
A 0.001 difference on a 0.004 total would be worth showing.
That was a bad example. Try 1,000,000 moving up to 1,000,069.
I'm sticking with relevance. A >25% rise is what we're talking about.
A 25% raise would show up with the y shits at zero. As would any significant increase.
True.