Charts and graphs are the building blocks of data visualization – and your dashboard. However, picking the right chart or graph to depict unique sets of data can be a challenge. The goal is to represent your data in a way that’s easy for users to grasp and digest, which involves two essential elements: aesthetics and function. In other words, the right chart or graph is both beautiful and useful.
In order to find the right visual aid to accompany a specific metric, you’ll need to ask yourself the following question: What am I doing with this data? Are you comparing sets of information? Do you need to break down a larger set of information into small, bite-sized pieces? Are you hoping to see how your data has changed over a given period of time? This guide is designed to help you learn more about the most commonly used charts types and help you find the perfect fit for telling your data’s story.
Some of the most commonly used charts and graphs are:
Using Charts and Graphs
Line and Area Charts – These graphs are valuable in situations where you want to understand changes or trends over time. For example, if you want to see how sales have gone up (or down) in the last year, a line chart might be your best option because it can provide visual aids for both value and time. Area charts offer similar flexibility with a slightly different aesthetic. They are particularly effective at showing cumulative impact and data composition over time.
A Closer Look at Line Graphs: Line graphs are one the most common and versatile chart types, so they deserve a little extra attention. There are three sub-types of line charts to discuss: trend, sparkline, and stacked.
- Trend lines are the most commonly used line graphs and indicate the tendency or “trend” of your data. If you have a graph depicting sales growth, a trend line could be used to show exactly how fast your sales are growing and, given the current sales rate, where they’ll be in the future.
- Sparklines, on the other hand, are best used for displaying a quick “summary” of information. If you want to convey a general idea about your data, such as increased customer retention rates or website traffic, you can do that with a sparkline. Because they are so tiny, sparklines are best used to show general trends – not specific numeric values. If you want users to have more information about the data in your sparkline, consider using a drilldown.
- Stacked line graphs are unique because, as the name suggest, they function like several line graphs stacked on top of each other. Unlike most line graphs, there is not opportunity for the lines to overlap. Much like a pie chart, a stacked line can be used to show the cumulative value of several smaller data sets. This way, users can compare data sets as part of a greater whole.
Bar Graphs and Column Charts – Column and bar charts are a great way to compare data sets. Let’s say you need to compare spending or revenue data for each department in your organization. A bar graph could help you evaluate the performance of each department by showing the highest and lowest values, along with everything in between.
Bar Graphs vs. Histograms: Bar graphs and histograms may look similar, but don’t be fooled! In fact, there are a few fundamental differences between the two (besides their visual similarities). Both are used to compare values, but histograms are better employed for comparing information across a continuous string of variables. A bar chart might provide the insight you need to compare sales by region, while a histogram could more adeptly show you the distribution of sales according to price point.Another important difference between histograms and bar graphs is the value represented by each “bar.” In a bar graph, each bar represents the value of an item. In a histogram, though, the bars delineate how many values fall into each bucket along the x-axis.
Pie, Funnel, and Pyramid Charts – Pie, funnel, and pyramid charts are dissimilar in their appearance but serve a common purpose: understanding how a larger data set breaks down into smaller pieces of information. If you wanted to know what percentage of sales came from referrals, cold calls, and other avenues, a pie chart would be useful. If you wanted a more detailed breakdown of each component and how they relate to one another, consider a funnel chart.
The Problem with Pie Charts: Despite their omnipresence in popular culture and PowerPoint presentations, pie charts have a less than sterling reputation. Edward Tufte, considered by some to be the “Jimi Hendrix of Data Visualization,” famously derided the pie chart, saying “the only worse design than a pie chart is several of them.” Still, the pie chart is a familiar and dynamic element of many data visualizations, and can be an effective chart if used properly. First, it’s important to recognize the drawbacks of this particular chart type.
- It’s more difficult for us to perceive differences in area vs. differences in length
- Angles in a pie chart can make it hard to see proportion
- Often, the colors used are incongruent and not friendly to the colorblind
How to avoid a pie chart fumble:
- Make sure your metrics add up to 100%! A pie chart must display a meaningful whole.
- Keep the center of the pie chart clear and defined. This is the best way for people to judge angles and proportion.
- Consider only using pie charts when showing part-whole relationships. Bar charts are better for showing relationships of parts to other parts.
- Never use (or eat) more than six slices of pie. Any more than that is too much to digest!
Speedometer, Bullet, and Thermometer Charts – If you want to visualize a single data set that has a threshold or an upper limit, like a goal or “danger” zone, a speedometer or a thermometer can be very effective. These chart types are best for representing numeric data that falls within a certain range. A speedometer will easily show where your data falls within that range, while a thermometer can show how close you are to your upper limits.
5 Use Cases for Speedometer, Bullet, and Thermometer Charts:
1. Comparing your current metrics to those of the previous period
2. Showing progress towards a goal
3. Comparing actual vs. target values
4. Visualizing an important threshold, such as a “danger zone”
5. Easily measuring whether your current numbers are good, bad, or satisfactory (helps to use red, yellow, and green colors in the background of these visualizations)
Tabular Charts – When you need to get granular, tabular charts are the way to go. These are best used as drilldowns, because they can be a bit overwhelming on the initial view. Nestled beneath another chart type, tabular charts give a curious audience all the detail they could ever want on a particular data set.
- When you need to look up individual values
- When precise numbers are important/required
- When both the summary and detail values are included
- When you need to compare individual values, but not an entire series
- When the information you want to communicate involves more than one unit of measure
How to Choose Your Charts
Choosing the right charts can help you create a dashboard that’s not only visually stunning, but also guides users on a narrative journey to insightful, actionable conclusions.
When it comes to picking the perfect chart to suit your data, there are four questions you need to ask. In reality, these four questions are the ones your data asks; you just have to listen. These four questions translate to the four major categories of charts and graphs in data visualization. Once you’re able understand which category is right for your data set, selecting the best graph is a piece of cake.
1. Distribution: Where do your values fall?
First up, distribution charts. This is the first and possibly most complex type of graph. Three of the most common types of distribution charts are:
- Stacked Ratio Areas
- Scatter Charts
Data best displayed in a distribution chart begs the following question: “Where do your values fall over a continuous set of data points?” Additionally, distribution charts take outliers into consideration and visualize them, making it easier for users to see where certain items fall outside the norm.
2. Relational: How do variables relate to each other?
Gantt charts, bubble charts, and treemaps demonstrate the relationship between given variables. Scatter charts also fall into the relational category. In other words, you can use them to demonstrate distribution or relationships – or both!
3. Composition: What parts make up the whole?
Composition charts answer the simple question: What parts (or percentages) make up the whole of the data? Like distribution and relationship graphs, there is some overlap with charts that fall into the composition category. Namely, treemaps and stacked area charts, which can also be found in the relationship and distribution buckets respectively.
The composition category has many, many charts in it which can be divided into two sub-categories: charts that show time and charts that are static.
Composition graphs that show data over time:
- Stacked Area
Composition charts that show static data:
- Pie Chart
- Ratio Bar / Column Charts
4. Comparison: How are these values similar or different?
You can divide charts in the comparison category into two sub-categories: graphs that compare values over time and graphs that compare values among items.
Of comparison charts, column charts, bar graphs, and trend lines are probably the most commonly used. However, if your data includes added layers of information, a stacked bar chart or cluster bar chart might be the right answer instead.
Other common graphs in this category include:
- Population Distribution Charts
- Horizontal Bullet Charts
- Metrics Charts
- Vertical Bullet Charts
- Tabular Charts
At their core, comparison charts are useful for showing pieces of information as they compare to each other and another metric, such as time or financial value. Additionally, comparison charts are useful for providing precise information because you can label them in detail without creating any visual clutter.
Need more help with chart selection? Check out this infographic.
Chart and Graph Best Practices
Most charts are versatile and relatively forgiving, but you can get the most out of them by following a few simple guidelines:
- Each time and value interval between marks on the graph should be consistent.
- You don’t always have to start the axis at zero. However, most people assume the axis will start there, so it’s generally a better idea to start at zero if you’re showing the chart to a broad audience. If you’re starting from another value, make sure that the axis labels are clear and legible.
- If on the X-axis, time should always run from left to right. Otherwise, you’re likely to confuse your audience.
- Labels are important, but you shouldn’t need to include guidelines or explanatory information – your chart should speak for itself. If you need to provide guidelines, consider finding a simpler way to present the data.
- While 3-D graphs might seem fun, they’re actually much more difficult to interpret accurately. Keep it flat and stick with 2-D for maximum legibility.