3 ms·

Also here graphs are confusing as they don't show the full y axis (please don't do that!) https://towardsdatascience.com/misleading-graphs-e86c8df8c5de https://

• mr_mitm 3y ago
  Disagree, at least when stated this generally. This article is explaining it thorougly: http://www.chadskelton.com/2018/06/bar-charts-should-always-start-at-zero.html http://www.chadskelton.com/2018/06/bar-charts-should-always-... (The article you linked is not accessible to non-medium users, by the way. Apologies if it covers caveats.) For bar charts it's a good rule of thumb. For line charts, not necessarily. Scroll down to the cases where it makes sense to zoom in. Imagine you plot the global average temperature of the last 200 years including the zero. You could barely see the changes, but they've been dramatic. Use Kelvin to make this effect even stronger. Which brings up another point: The zero is sometimes arbitrary. If instead of a quantity you only plot its difference to some baseline, all that is changing is the numbers of the y-axis, but the actual plot stays the same. Is it now less misleading? I say no, because the reader must look at the axes and the title either way. So please do zoom in to frame the data neatly if it makes sense to do so.
• causal 3y ago
  For headline news that's probably a good rule of thumb. But in an academic context, especially data science, anyone who is data literate should know to check the axes.