That seems a rather arbitrary trait to dismiss a researcher for. It is possible to be correct with flawed design choices.

Honestly communicating your results is just as important as getting good ones, and a junk chart like that is at best gratuitously confusing.

I'd say everyone should have to read Edward Tufte, but everyone has and it didn't much help. 90% of the Q&A at his seminars is people asking questions that are variants of "but I'm used to doing it the old bad way, so isn't that really just as good?"[1] It was mildly amusing watching Mr. Tufte grow increasingly exasperated realizing just how few of his students had learned anything from the experience. I imagine that was a big part of why he retired to become a sculptor.

[1] My personal pet peeve is the obsession designers appear to have with ensuring that there are no visible indications of which part of their design is meant to be clicked on.

that pie chart to the left is not poor design choice. it demonstrates actual incompetency.

I didn't get it at first but you're right, it's horrifying. They put the total and the breakdown into the same pie chart.

Yup. Something I wouldn't consider adequate work from a 6th grader.

Oh man. I was annoyed by the breakout of the second pie chart which seemed needlessly confusing. But the left one is worse for double counting everything.

And it has a typo! ("Analys"[sic]). That doesn't really bode well for how much rigor the rest of paper had, either in its creation or review(s).

Why? The labels are given in absolute units but they are correct.

The issue is that the pie chart included a slice for the total and then two slices for each of the two components that sum to the total.

The pie totals 200%.

One of the labels has a typo.

Yes, you are of course right. One shouldn't dismiss good work because of some flaws. But anyway... this is a rather silly pie chart and I wonder how this was able to pass through multiple authors and peer-review.

The popular conception of science is really outdated in this respect. Peer review still matters, but most meaningful peer review happens post-publication, not pre-publication. People should not take published papers as reliable statements of "what the science says".

Scientists are not going to expend a lot of effort on the thankless work of peer review, taking time away from their own careers, purely out of professional integrity and the goodness of their hearts. And their competency to do so is uneven at best anyway. What happens when the people who wrote this paper review other people's data?

Can ANYONE actually say what is wrong with it? I see no issue other than ascetically it could be better but it shows the data correctly.

You would probably never go to a doctor if you go to a medical conference. The number of red spelling lines from word you see on graphs would make you throw up lol

Is that because the "Consumables ingested" on the right side looks like Pac-Man eating the other two categories?

The pie chart on the left double counts everything. The total is represented as a slice in the pie chart.

Probably a case of automatically converting an Excel column into a pie chart.

That's what came to mind. And they simply selected a pie chart instead of a bar chart, and it wasn't super wrong because they were trying to show subcomponents of a larger value.

Meh. Not worth the HN thread. It's like picking apart why someone wrote code in a suboptimal/messy way when in reality they just wanted to get something done under a constraint, and the author would totally agree if they were here to defend themselves.

Yet here are people saying the author can never be taken seriously because of it.

I didn't think anything of it at first but that is indeed a strange design. Especially since the right part is a breakdown of one of the pieces.

You might care to state why the chart is problematic.

"Total number identified" is one of the pie slices, the other slices are subsets of this slice.

Fair enough, thanks.

The charts themselves would be better expressed as a Sankey flow, perhaps.