Optimize Your Dashboard: How to Choose the Right Chart or Graph

Contributed Post By,  Naill McLean, product manager, insightsoftware.com

As a product manager for a business performance management company, I commonly get asked, “what types of chart should I use to best represent my data” or “when should I use a specific chart type?” Some of this comes down to personal choice, but as a rule, if you need to explain it later, you chose wrong. In this blog, I explain why you would want to use a chart, give examples of when to use them, and include a few best practice tips.

Why do we use charts and graphs?

Let’s first consider why people even use charts and graphs in the first place.

Every day, we should be making decisions that will positively impact the performance of our organization. In the traditional list report, it can be difficult and time-consuming to pick out the information that will steer your business in the right direction. If a report takes an hour to analyze or gives you a headache due to the number of problems it describes, you will avoid using it or miss information. By using a visual representation, you can make decisions faster and respond in a way that will drive business performance.

Let’s take an example from everyday life. Most likely you drive. Even if you don’t, you know about a car’s dashboard. A dashboard is great for visualizations. Typically features are in the form of a one-dimensional graph: a fuel gauge, speedometer, or oil warning light. These are all things you can glance at and know if all is well or if you need to adjust your driving. Could you imagine if you had a report running with lots of text telling you all the same information? You would crash before you even found what you were looking for. It’s the same for business. Charts should, at a glance, let you know everything is okay or if you need to “adjust your driving.”

As an example, check out the following dashboard. On the right-hand side, there’s a traditional income statement. And, straight away, I am sure you glanced to specific numbers and ignored most of what was in front of you. This income statement includes three months of data, but, can you tell from looking at it whether revenue is good or bad? We don’t know, because without comparing it to an expectation, we don’t have enough information. I could add last year’s number or a budget number, but then we would need another three or six columns and the report becomes even more complicated.

Now let’s take the top left graph. This shows five months of data with the revenue, cost of goods sold and the gross profit. You can quickly see that, although revenues have been increasing, they are not as exciting as they were on the text report because the revenue is returning to what we had in February.

Adding another layer – the metric on the bottom left shows that revenue is close to the budget. This chart shows that our data is no longer the good news we initially thought.

Let’s consider which charts to use

Now that you’ve seen how charts can show relevant data more efficiently, and that a list report often shows too much irrelevant data or tries to solve too many problems, let’s discuss which charts you should use.

I recommend first identifying the dimensions of the data you need. This quickly reduces the number of chart options. For example, if you have a count of orders by month or a count of orders by salesman, this is one dimension. If you have a count of orders by month and by salesman, then you’ll need two dimensions. Knowing this, you’ll realize that you don’t want to try and show two dimensions on a pie chart, as it would make for a lot of pies that are difficult to compare.

One-dimensional visualizations

Pie charts are best when you are trying to compare parts of a whole.

Best practices

  • Don’t compare multiple pies against each other, e.g. 2015 vs. 2016, as it will be difficult to interpret. A 100 percent stacked column would be better.
  • Don’t add too many segments as it will become unreadable and cluttered. Use six to seven max.
  • Order slices by size – largest first going in a clockwise direction.

Speedometers offer a spectacular visualization to present performance. They’re normally used on a dashboard. At a glance, a consumer can see how well something is performing.

Best practices

  • Best suited for data that is changing in real time or near real time.
  • Don’t add multiple categories as many needles would be confusing.
  • Show indicators of what is good and bad.

A funnel chart shows a series of steps and the completion rate for each step. This can be used to track the sales process or the conversion rate across a series of pages or steps.

Best practices

  • The size of each section should reflect the size of the data set to provide quick analysis.
  • The funnel is suited to a business process that has steps that feed down from one to another.

Waterfall charts help show the cumulative effects of sequentially-introduced positive or negative values. The columns are color coordinated to distinguish between positive and negative values with the final column showing a total. Waterfall charts can be used for various types of quantitative analysis, ranging from inventory analysis to performance analysis.

Best practices

  • Use contrasting colors to highlight differences in data sets.
  • The final total column should be a different color as this will usually be calculated from the previous columns.

Metrics graphs are often used to show KPIs or progress towards a goal. They allow you to quickly know when to act. Much like the warning light on the car dashboard, these are a great tool to drive action. They are a personal favorite of mine because they can be very specific and easy to understand.

Best practices

  • Use a visualization that has meaning, as this will help with user adoption levels.
  • Always have a target and a plan of what actions should be taken if the target is not being met.

 

 

 

 

 

 

 

 

Bullet graphs would typically fall under the metric section as they have the same characteristics. I’ve labeled them separately, however, as they often include milestones.

Best practices

  • Always have a target and a plan of what actions should be taken if the target is not being met.
  • Set milestones for clear visibility of progress towards the target.

Two-Dimensional Visualizations

Line/spline charts are used to track changes over short and long periods of time. When smaller changes exist, line graphs are better to use than bar graphs. Line graphs can also be used to compare changes over the same time period for more than one group.

Spline charts are specialized forms of conventional line charts. Unlike conventional charts that connect data points with straight lines, a spline chart draws a fitted curve through the data points. These feel more modern and are more aesthetically pleasing. This won’t impact the results, of course, but it can help with user adoption.

Best practices

  • Don’t add lots of lines to a chart, they will become difficult to identify from one another.
  • Change the start point of the y-axis so you don’t have lots of white space at the bottom.

Area charts, area spline charts and stacked area charts are all line charts but with the area between the lines filled in. The stacked version can be useful to show how many categories can add together to show the makeup of a total.

Best practices

  • The filled areas should be transparent so information is not hidden by another fill.
  • Don’t add lots of categories, as this will cause clutter.
  • It’s not advisable to completely stack area charts as the largest area will often look like a background.

Bar charts and column charts display data using rectangular bars where the length of the bar is proportional to the data value. Both are used to compare two or more values. However, their difference lies in their orientation. A bar chart is oriented horizontally whereas the column chart is oriented vertically. Bar charts are best suited to times when the label is longer for the categories.

Best practices

  • Use a column chart for sets of data that have 12 or fewer categories.
  • Column charts should be used if there are negative values.
  • Use a bar chart if there are more than 12 categories.
  • Use a bar chart if the labels are longer.
  • Start from zero in the value axis.

Stacked bar charts and stacked column charts are used to compare categories between different groups or to track changes over time. They are particularly suited to categories with long names. However, when trying to measure change over time, bar graphs are best when the changes are larger.

A stacked bar chart will combine multiple categories with a common Y-axis dimension. Stacking is useful to be able to see the composition of a total value for a dimension.

Best practices

  • Space bars out appropriately in these charts.
  • Make sure to use consistent colors.
  • Start from zero on the x-axis

100% stacked bar or stacked column charts allow you to compare category breakdown as a percentage. These can be useful for things like product mixes compared across cities.

Best practices

  • Needs to be accompanied by another chart showing the total context or the total need to be added as a point on the chart using a dual axis.
  • Interactive charts can be a real improvement as these allow users to see what would happen if a product was removed from the mix.

Scatter charts are useful to show the relationship between data and highlight categories that fall outside of the norm. This can be helpful when trying to identify the outliers in a process and highlight that further investigation is required.

Best practices

  • Don’t add a lot of categories; more than five will become cluttered.
  • Add an average line to easily highlight the points that deviate from the norm.

Dual-axis charts can help spot a trend between numbers that are normally too far apart to put on the same chart. In the following example, we have a large monetary value and small counts on the y-axes and supplier on the x-axis as a shared axis. This allows us to see the relationship between data that has a shared axis. In this example, we can see that the DHL supplier is sending much larger quantities of invoices than our other top suppliers. This would indicate an increased overhead that could be addressed with the supplier.

Best practices

  • The data must have a correlation or show the lack thereof between the data sets.
  • The left axis should be used as the primary category in countries where the user will read left to right.

 

Conclusion

Hopefully you now have a better idea of what charts are available, how to narrow down your choices and some best practices for implementing your chart.

What dos and don’ts have you learned while creating charts for your own reports?

 

Leave a Reply

Your email address will not be published. Required fields are marked *