Best Practices for Creating Useful Dashboards | Hacker News Reader

Best Practices for Creating Useful Dashboards

dataschool.com

The chart type picture is somewhat small resolution (at least thats how it looks on a 4k screen).

There's a recommendation of a pie chart in there - stopped reading there. /s Seriously, though, even when it comes to a distribution to a composition, I'd go with a (vertical) bar chart - there's plenty of literature/comments against the use of pie charts.

"The only thing worse than a pie chart... are two pie charts"

I'm not sure if I missed it whilst glancing over the pages, but I didn't see any mention of interactivity. I think that especially nowadays, being able to slice-and-dice is quite important for a dashboard. There's so many angles from which the data can be looked at.

"being able to slice-and-dice is quite important for a dashboard. There's so many angles from which the data can be looked at"

You make the fundamental mistake of confusing a dashboard with an analytics tool. The former must function without interaction because it must deliver its value without the user working for it. The minute they have to "slice and dice" to get an answer then it's not a dashboard. Dashboards exist (as their name implies) to be glanced at in order to establish whether any action should be taken.

Hi, thanks for the feedback here. We mentioned the interactive display piece in the section on sharing the dashboard. Interactivity is important for a dashboard, especially in this day and age.

I noticed you didn't reply to the pie-chart feedback. I'll second that criticism: pie charts are unwise.

Agree, pie charts may have a narrow, if any, use case, however, some people LOVE them. Maybe not as much as I love commas, apparently.

Artists like pie charts because they're colorful and break the monotony of right angles. Scientists dislike pie charts because they don't convey information well. Who is your audience?

I totally understand the feeling/frustration of pie charts being overused and often where they shouldn't be, and for the most part, I agree. I think it's always a good idea to stop and ask, "can a different chart convey the information better?" And the answer will almost always be "absolutely" when looking at a pie chart, no denying that. I will throw these two cases out for consideration of where a pie chart may be acceptable though:

1) As you mentioned, they're visually pleasing. No, that doesn't make up for when it's hard to read data, but I think a case can be made for snagging people's attention and getting eyeballs on a report you send out. If a pie chart is the hook that gets people interested in your data, maybe it's a sacrifice worth making. Artists aren't the only people who can find things visually appealing.

2) Basically, storytelling. There's no rule that says you can't show the same/similar data in different ways to help tell a good story. In some situations, it could make sense to start with a high-level visual breakdown of a group, and then use more detailed charts to get into the details you want to focus on (like how different segments relate). Or maybe the story I'm trying to tell has a ton of sub-groups which all pale in comparison to one dominant one. I think a pie chart visualizes that situation (where exact details aren't important) pretty well.

I ultimately just don't think it's a binary thing, where only artists will enjoy having a little visual variety and that any and every pie chart is a 100% waste of everyone's time. Most of the time, they're silly and an excuse for different shapes and colors, but I like to think even pie charts have their place in data visualization.

Site appears to be experience the hug of death, Google cache:

https://webcache.googleusercontent.com/search?q=cache:rCmcbq...

Thanks for the heads up here. We just took a look and things seem to be operating normal now.

Voiceover: It wasn't.

Thanks for the write up!

Side note,

Adding this css would help with scrolling on mobile.

  overflow-y: scroll;
  -webkit-overflow-scrolling: touch;

Nice, thanks. I'm still figuring out this WP plugin we're using for the guides, I'll give this a shot. Much appreciated.

While this is good in the sense that 99% of dashboards are worthless drek and anything that helps that is good, the examples given here are surprisingly poor. Even a cursory reading of the book they themselves recommend (Stephen Few's "Information Dashboard Design") would know this.

In fact, it appears the author may not have even read Few's book because they fail to mention some fundamental points about dashboard design that need to be understood if the design isn't going to just be a form of simple data analysis tool.

Overall, if you're interested in dashboard design, I'd give this course a miss and just read Stephen Few's book on the subject.

My teammate, Tim, put together this guide of tips for creating a simple and useful dashboard. We've been working on this new site/team putting together helpful articles for business users to learn about using data for work. Comments welcome!

As a long time dashboard designer, I love this site

https://www.data-to-viz.com/

Best Practices or "Building Dashboards for Dummies?"

The guide is meant to be fairly "intro-level". Did you feel the points AREN'T best practices, or where you hoping for more information?

Boo. No pie charts.

Never ever?

Almost never, and only when you know exactly what you are doing: https://eagereyes.org/pie-charts

Reply on news.ycombinator.com