The Data Point #2: 5 Best Practices for Useful Dashboards

Read time: 3 minutes

The beauty of data visualization is the ability to reveal problems or solutions we didn't realize were there.But charts can either add to or detract from your overall goal.

There’s more to analytics than just visualization, but a good dashboard can contribute greatly to helpful insights. Conversely, a bad dashboard will make the data even fuzzier.

We’re going to reference the dashboard below throughout the article. This is a bad dashboard.

The colors look pretty at first glance, but I’ll explain why this is a bad example.

There are 5 best practices in dashboard creation that I find especially helpful, whether you’re using Tableau, Power BI, or Excel to create your dashboards.

1 | Every chart should serve the end goal of the analysis.

It can be tempting to want to feature as many charts as possible in our dashboards. Sometimes we also want to add a lot of variety as we think it will add more perspectives to our analysis.

But this often works against us. More charts don’t always equal more insights. Sometimes the opposite is true.

I'd rather have 3 meaningful charts over 6 that all say the same thing.

You also want to make sure that you’re choosing the best chart for what you’re trying to communicate. Bubble charts look cool, but is it the best tool for the job? Would a bar chart work better?

Take a look at the example I shared. There are just WAY too many charts to create any meaningful analysis and some of them are redundant. Look at the two “Sales by State or Province” charts. They say the same thing. The pie chart is especially useless here as there are way too many values.

More charts can create clutter, and clutter makes it difficult to draw insights from our dashboards.

Speaking of clutter. Let’s talk about blank space.

2 | Blank space is your friend.

As mentioned, in the first point, it can be tempting to want to maximize our visualizations with as much information as possible.

But a dashboard that is difficult to interpret quickly is not effective.

This is where blank space comes in.

Effective use of blank space is easier on the eyes and makes each chart stand out more.

Our example above has close to no blank space. Again, way too cluttered.

Think of it this way, say you’re going to a friend’s house. Which would you prefer: a house with dishes everywhere, stacks of newspaper, boxes, and trinkets filling up every corner; or would you rather a house that is neat, tidy, and organized? Exactly.

Less is often more. The same applies to dashboards.

3 | Create interactivity to boost user engagement and expand the analysis.

People like a dashboard they can interact with.

Adding dynamic filtering and tooltips that add more insights not only boost engagement, but they’re fun to use.

While some situations may call for a static dashboard, it’s typically more fun to play around with a visualization rather than just staring at it.

And as mentioned, the dynamic element can add more observations than can initially be seen.

Our example does appear to have interactive elements to it, but this point speaks for itself. Just be sure to add this to your charts whenever appropriate!

 4| Consistent coloring.

Colors should generally represent the same categories in each chart.

In our example, notice how teal represents “Office machines” in the treemap, but then also used to represent “sales” in the bar chart right below it. Confusing.

Speaking of the treemap, there are way too many colors. A bar chart using only one color would work better. The size of the bars would tell you what you want to know. Different colors add no value here.

Pay attention to your use of color. Keep it clean. Sometimes a few is all you need, but it always depends on the dashboard as a whole.

5 | The 5-second rule.

The 5-second rule states that you should be able to draw primary insights from a dashboard within 5 seconds of looking at it.

I use this strategy often.

If you’re having to focus too much on the details of a dashboard to understand the key ideas, then you should consider tapering it down a bit.

This mainly applies to big-picture ideas of your visualization.

I think I had to stare at our example for a whole minute before I could figure out what it was trying to tell me.

I should be able to understand what I need to know within seconds of looking at it.

Good example

Let’s wrap things up with a good and simple dashboard example. Here’s a call center KPI dashboard.

This one ticks all the boxes for a good dashboard in my book.

1. Simple. Just a few charts and some core KPIs.

2. Good use of blank space. The dashboard can breathe.

3. Interactivity. Notice the filters at the top.

4. Consistent coloring. One color: orange. The charts speak for themselves. Beautiful.

5. 5-second rule met. Right away I could see that Monday is the best day and mid-morning is the peak time.

I hope these 5 ideas helped to read about!

See you next week 👋

PS - I just published my new personal website! This will be a spot where I can keep all my resources in one place to help people begin or develop a career in data.

Check it out here

Whenever you’re ready, there are 2 resources that can help you:

1. If you’re looking to create a data portfolio but aren’t sure where to start, I’d recommend the following book:

   1. The Data Portfolio Guidebook: Learn how to think like an analyst, develop a project portfolio and LinkedIn profile, and prepare for the job hunt. Join 350+ learners here.

2. For help navigating the data job hunt, consider booking a 1:1 call with me. I can help in the following ways:

   1. Resume, LinkedIn profile, or portfolio review

   2. Career guidance

   3. Interview prep

      1. Schedule a time here.