Hacker News Reader

Comment by arscan | Hacker News Reader

Ah, that chart on the bottom is on a log scale. I wasn't expecting that... and if I didn't look closer I would have completely misinterpreted it (for example, it appears that the DoD is mostly shut down).

View on HN

It's downright deceitful, in my opinion. The entire page is trying to exaggerate the significance of the shutdown.

I personally think the author did this because otherwise the DoD would so heavily dwarf everything else that you wouldn't be able to see whats going on. But I agree, intended or not, this has the effect of heavily exaggerating the proportion of people on furlough. Its like using a log scale on a pie chart. It really makes no sense.

The author really should consider this and think hard about what point he is trying to make with that chart on the bottom.

It's not attempting to be deceitful. The problem is the DOD has 8x as many employees as the next largest agency. If spaced without a logarithmic scale, the rest of the agencies are so insignificant it doesn't have any effect at all.

To try to avoid the confusion, I'm working on another graph of percentages working for each agency, and a disclaimer about the scaling.

Mark more scale points: I suggest {1,2,4,6,8}x10^n for each n. That way the visual texture of the scale lines will immediately suggest "log scale".

Make a 2nd chart, with little inset zoomy lines to connect the 2 charts (like you see on a paper map that shows a zoom of the downtown core)

Like this: http://xkcd.com/radiation/

Is there a good js library for doing this style chart?

It seems like you're loading it with the Sean Hannity inane view that its basically no big deal though, right?

Exaggeration seems to be a common theme for one of the authors: https://twitter.com/JustineTunney/status/385843326085132289

Exaggeration was not the intent. We've updated our graphs to be more clear.

I agree that the bar chart is misleading, especially horizontal scale. Pie chart can show which department is being affected the most.

http://vida.io/discussion/WgBMc4zDWF7YpqXGR

NASA, HUD, ED are the most affected, not DOD.

We incorporated this graph with the permission of the author.

It just doesn't work. Similar visual comparisons between agencies mean vastly different things depending on their position on the graph. This is an example of a dataset that would have been better served by a simple list of numbers and percentages.

It took me a sec too. It's a confusing application, since the intent is to compare to adjacent rows.. the first one for example is "half furloughed" but only a slightly smaller red line. Visually confusing.

I like it. You need only look at a few log charts to get the hang of it, and if it wasn't a log chart, it would be hard to see the difference between organizations with less than 20,000 employees, because the bars would all be very short.

Edit: took another look, and it is quite a bit more confusing than I thought, because of furloughed vs total employees. I can't quickly tell if the furloughed for an agency is linear based on the percentage of the total employees for that agency, or is also logarithmic.

It's also logarithmic.

It should also be a stacked bar chart, not two individual bars.

Reply on news.ycombinator.com