Data Visualization enables you to easily draw insights from data and make better decisions. It is important to choose the right data visualization. If you don’t use the right data visualization to display your data, it can confuse the viewer, instead of educating him. Here’s an example of what happens if you don’t choose the right data visualization. Let’s look at the 2 charts below:
The sales trend is obvious in the second chart. In the first chart, you end up wasting time between legend and the pie slices.
How to Choose the Right Data Visualization
In order to choose the right data visualization, you need to know the popular chart types first. Let’s look at the 7 most commonly used data visualizations and how to choose the right data visualization
1. Number Chart
When to use number chart
A Number chart is a ticker that shows the real-time value of a KPI metric. It provides at-a-glance view of any metric such as total sales, number of signups, daily revenue, etc. It is the easiest chart to build; the only thing to consider is the time period of data you want to track. Do you want to show the number about past quarter, past month, past week, or just today’s value? It’s important to utilize the chart title to clearly label what the number stands for, to avoid any confusion. Also, it is advisable to place number charts at the top of a dashboard for easy visibility.
You can add a growth indicator to compare the number to a previous period and tell you if it has grown or fallen. You can also apply a threshold to visually alert you (by changing to green/red color) in case the number goes above/below the threshold.
Mistakes to avoid
Number charts are the first things people see and analyze on a dashboard. People tend to look at the numbers, check if there is any deviation from the norm, and drill-down the subsequent charts for more in-depth information. Number charts shout-out to the viewer, “Look here, this where you can begin analyzing your data”. So, if you have too many number charts on your dashboard, it can confuse the viewer and dilute your narrative. Limit the amount of number charts so that there is room for other visualizations such as line charts, or bar charts that can be used to provide more detailed information about the numbers.
2. Line Chart
When to use
Line charts are used to show trends, growth, or volatility, generally over time. They are used to find how one or more variables(y-axis) change due to another variable(x-axis). Line charts make it easy to spot trends, such as when does your sales grow, when it falls. For example, in the above chart, you can clearly see that the sales generated due to Email marketing are much more than other channels. Line charts also help you spot periodicity – e.g, sales peak every summer. They even help you figure out what trend is normally followed, and quickly spot any deviations.
Another great thing about line charts, is that you can combine it with other charts, such as bar graph. You can use 2 Y-axes – one for bar graph and one for line chart to show trends of 2 variables in a single chart.
Mistakes to avoid
Too many lines can make it difficult for viewers to understand your chart. They will need to constantly refer to the chart legend, to understand what they’re looking at.
Also, setup your y-axis scale to be close to the highest point. For example, in the first line chart, if we had set the y-axis to track all the way upto 20,000 (when max value is 7,000) our chart would have been squeezed and difficult to read. The top half of the chart would be wasted space and the data would look crammed.
3. Bar Chart
When to use
Bar Charts are used to show chronological data such as monthly sales numbers, or compare a variable across categories. The most useful part about a bar chart is that you can stack them side-by-side and immediately find out which one is bigger than the other. The chart below compares quarterly sales numbers. The single color here indicates that all bars belong to the same variable,i.e., sales amount.
The bar chart below compares complaint resolution time across multiple categories – email, telephone & social media. In this case, the multiple colors clearly indicate which bars stand for email, telephone & social media. This makes it easy to compare multiple channels for a single value of x-axis, or compare single channel across multiple values of x-axis.
Mistakes to avoid
Since time is best expressed left to right, it’s better to show time in bar chart chart, from left to right. Also, like many charts, when you have too many values, a bar graph quickly becomes cluttered.
4. Column Chart
When to use column chart
Horizontal Column charts are great for comparative ranking, like a top-ten list. They can also be used if your data labels are very long. However, it’s advisable to list them in an order that makes sense. Either arrange them by value or choose another logic for the labels, like showing them alphabetically(like we did above). The length of the columns makes it easy to rank them, find out who is the top performer.
What to avoid
One of the most mistakes people make is to confuse this chart with bar graph, and use it to plot trends over time. Being horizontal, it becomes a little difficult to spot time-based trends, such as growth over time, on this chart. Also, too many columns can make it difficult to compare values. Just like line charts, it is important that you scale the x-axis very near to the highest value. Else the columns get shortened and a lot of space on the right will be wasted.
5. Stacked Charts
Stacked charts are the best to show part-to-whole relationships. That’s when you’re comparing data to itself and show it as a percent of the total. In the example, the total length of sales cycle every month is not as important as the fact that most of the time is spent during product trial (52% in Nov’14, 63% in Jan’15, and 54% in Mar’15).
Pie charts are the easiest way to show a single part-to-whole relationship. For example, the information in the first bar can be plotted on a pie chart. But, if you want to show same information over different time periods or other categories, then stacked bar charts are much more versatile.
Mistakes to avoid
When you have many data points, columns become quite thin and leaves hardly any room to label the chart. To make your chart intuitive & enjoyable, use good colors, enough spacing and a balanced layout.
6. Pie Chart
When to use
As mentioned above, pie chart is the simplest way to show a single part-to-whole relationship. However, it is advisable NOT to use it under any circumstances, simply because it can be difficult to look at pie slices of similar size and determine which one is bigger. You will end up reading pie labels, which defeats the whole purpose of this visualization. Also, as the number of slices increase, even data labels get squished and become difficult to read. For, example, in the following pie chart, can you determine whether revenue from Fax or Mobile is higher – without using the labels?
It’s always easier to understand the same data by plotting it on a bar chart or a column chart.
7. Scatter Plot
When to use Scatter Plots
Scatter plot is a great way to bring out the correlation between 2 variables, in a large data set. The data should consist of value pairs – one for the dependent variable and another for the independent variable. Scatter plot shows how the independent variable (plotted on x axis) impacts the dependent variable (plotted on y axis). When the data is plotted, the correlation becomes evident. Adding a trend line helps you show the correlation and indicate its statistical significance.
What to avoid
Scatter plots work only when you have lots of data points, and you’re specifically showing a correlation. If you have only a few data points, a scatter plot will not provide much insights. It’s valuable only when there are enough data points to see clear patterns.
The first step towards successfully visualizing your data is to ask the right questions. Some simple questions to help you get started with data visualization are:
- What KPIs do you want to display?
- What kind of data do these KPIs contain – comparison, trend, relationship, etc?
- Who is the target audience? Executives look for high-level overview, managers & analysts look for detailed information
- How will you target audience use the information? Executives use it make decisions, analysts use it to get answers, explain trends & drill-down to get more details
Answering these questions before you design your graphs & visualizations will help you figure out which KPIs to show and choose the right data visualization.
Finally, here’s a handy diagram to help you choose the right data visualization to use in your dashboards:
When designed well, data visualizations can serve as an indispensable tool for gauging performance and making decisions. On the other hand, if you don’t choose the right data visualization then it can cause confusion, waste time and even misguide users.