Business leaders depend on data to make decisions. In order to act quickly, they don’t have time to sift through spreadsheets or query massive databases. Instead, they rely on data visualizations—charts and graphs that provide a pictorial summary of a dataset to show business leaders what they need to know.
Today’s businesses increasingly rely on data analysts to examine datasets and create data visualizations that aid decision-makers in various roles. According to Alice Mello, a professor in Northeastern’s Master of Professional Studies in Analytics program, these professionals should refine two major data visualization competencies: the ability to work with datasets and an understanding of the best ways to clearly depict the conclusions drawn from looking at the data.
Keep reading to learn how to improve your data visualization skills and the importance of effectively communicating your findings to a range of audiences.
Download Our Free Guide to Advancing Your Analytics Career
A guide to what you need to know, from the industry’s most popular positions to today’s sought-after data skills.
Key Steps and Skills of the Data Visualization Process
Step 1: Working with Datasets
Since good data is the foundation of a good data visualization, it’s important to first understand the data set that’s being used to create the visualization, Mello says. “You need to be able to conduct all the exploratory data analysis that’s necessary to see what stands out,” she says. “You need to see the patterns and the ‘a-ha’ moments in order to tell a story.”
These patterns could include a decrease in sales at a certain time of year, increased production from one particular factory, different patterns in employee computer usage since a department shifted to remote work, or just about anything that might be of interest to a business leader. The value of data visualization is presenting this data in a way that helps business leaders extract meaningful information at a quick glance and without the need for further explanation or analysis.
A background in computer science or engineering is certainly helpful, Mello notes. Still, professionals from fields such as economics, finance, or marketing who have taken courses in statistics should have the basic data management skills necessary for creating data visualizations. Knowledge of the R programming language is also one of the core skills for data analysts who often create visualizations, as this is the programming language commonly used for large sets of data and for running predictive analytics.
There are three critical skills for working with a dataset before it can be used to create a visualization: Understanding how to manage databases, learning how to use data visualization software, and knowing how different audiences may use the data. A cleaner dataset will enable more accurate visualizations and ensure that the work can be done in less time.
Learn how to manage databases.
Managing databases and being able to retrieve data are some of the most valuable data visualization skills, Mello says. This can include tasks such as properly naming columns in a database, quickly searching a database, and joining tables. It’s also helpful to know how to run statistical tests and import data into dashboards, she adds.
It’s also important to know what types of data are represented within the database, according to a tutorial from the data management professional organization ISACA. Are the variables fixed data points, such as a specific city or state, or are they ranges, such as numbers between 51 and 100? From there, you can determine key relationships among the variables, which could include growth over time, a set of rankings, or deviation from the norm.
Become proficient with data visualization software.
Professionals should have a good grasp of the software tools that allow users to import data sets to create visualizations. Tableau is widely used among medium-sized and large companies, and other options include Domo and Microsoft Power BI. Mello notes that these products focus on creating data visualizations and not managing data sets, which is why it’s important to know how to work with data before it’s time to put it into charts and graphs.
Understand the data’s audience and purpose.
While examining a dataset that will be used to make visualization, it’s important to understand how the intended audience is going to use that information. Data scientists, for example, will likely examine the dataset differently than sales managers or business analysts. What’s more, data scientists may need to see much more of the dataset than other end users in order to draw a conclusion.
Knowing your audience in advance will help you use the right data to create the most effective visualization. This requires getting to know your audience through stakeholder meetings or focus groups to understand what data they use to make decisions and what additional data they need. This will ensure that the right data goes into a visualization, and that different needs for different audiences will be addressed.
Step 2: Creating Visualizations
Once you have examined your dataset, determined which variables you’d like to compare, and prepared the data to be imported into a data visualization tool, it’s time to create the visualization.
The goal should be to tell a story with data, Mello says. Think of the structure of a narrative—there’s background information, there’s the revealing of your critical findings, and there’s the discussion of what those findings mean for the future. “You need to make it meaningful for the audience that’s going to see it,” she says. “How do you put the story together? There’s conflict and resolution. Try to walk them through that.”
Key tactical skills for this process include choosing the right type of visualization, keeping the visualization simple, and ensuring that it’s easy for an audience to understand.
Choose the right visualization.
While there are seemingly countless options for depicting data in a visualization, ISACA’s tutorial breaks the major types of visualizations into four categories, based on how viewers are meant to look at the data:
- Comparison of variables, typically through a line graph or bar chart
- Distribution of variables, using a histogram for many variables and a chart for two or three variables
- Composition of variables, with column or area charts to show changes over time and simpler charts for static variables
- Relationship among variables, typically done with scatter or bubble plots
Other examples of visualizations include color-coded maps, heat maps, and box plots. It’s also possible to combine multiple types of visualizations into a single picture. For example, a bubble plot on top of a map can show data such as population density by state or county. A combination of a bar graph and line graph can show static variables (e.g., quarterly sales figures) and changing variables (e.g., the growth in quarterly sales over time).
Make it easy to read.
Design elements such as font, color, line thickness, and data arrangement can all impact a data visualization’s readability. This matters not just for practical reasons but also for matters of accessibility, as visualizations should be readable for the colorblind and the visually impaired.
ISACA offers several suggestions.
- If the goal of a visualization is to compare a set of variables, list the variables from largest to smallest and make each variable the same color. This will focus the eye on the items that need to be compared.
- Ensure there’s enough contrast between foreground and background colors.
- Avoid combinations of colors that are difficult to differentiate (such as orange and yellow) or combinations that the colorblind cannot differentiate (such as red and green.)
- Choose color combinations that can be easily replicated in a black-and-white visualization.
- Use different line styles, such as dotted, dashed, our double lines, to distinguish patterns if necessary.
Data analytics consultancy Key2Market provides additional recommendations.
- Use a single font with clear variations such as bold text and larger/sizes instead of multiple fonts.
- Align and sort all design elements to the left side, as that’s where a viewer starts looking at a page.
- Place labels for variables as close as possible to a chart’s bars or a graph’s lines to make it easier for the eye to link a label to its corresponding variable.
Keep the visualization clean.
ISACA’s tutorial points out that our brains have a limited cognitive load. If there’s too much information in a data visualization, then the brain cannot make sense of the data. Anything that doesn’t support the message of the visualization is simply going to distract viewers.
“You should have a clean dashboard that doesn’t include too much information,” Mello says. The size of your dataset may tempt you to create a graph or chart with many variables, but the most effective data visualizations focus specifically on the information that matters.
As a Harvard Business Review tutorial notes, viewers typically only have a few seconds to look at a chart, interpret it, and take action. For example, a line graph that tracks too many variables or charts a variable across a longer time period than necessary will make it difficult to determine where viewers should focus their attention. Think about the specific point that you’re trying to make with your data—the change in prices over time, for example—and get rid of anything in the visualization that is a distraction from that point.
Step 3: Communicating the Data’s Implications
While the goal of a good data visualization is that it can speak for itself, in many cases, you may be called upon to discuss your visualization in front of a stakeholder audience.
Reports prepared for business leaders are often the focal point of internal meetings, and you may be tasked with creating a slide deck and a short oral presentation for the group. If you aren’t leading the presentation, you may be asked to support the business leader by providing additional details or answering follow-up questions.
In some cases, data visualizations are used for external purposes, ranging from a shareholder presentation to marketing collateral such as an infographic, social media post, or white paper. Here, you may be asked to write about the data visualization—in something as short as a 280-character tweet or as long as a research report—or to provide a condensed version of a presentation given to internal audiences.
Use clear, concise language.
Regardless of the scenario, data visualization expert Bill Shander suggests that written and verbal communication about data visualizations works best when it uses clear and concise language. Being able to explain your work in simple terms is one of the most critical data visualization skills, as it ensures that a range of stakeholders and audiences can understand the data and what it means for their work and in their lives.
Build Your Data Visualization Skills
The Master of Professional Studies in Analytics program at Northeastern University helps analytics professionals prepare for a career using their skills in extracting, translating, and visualizing data to help organizations make both tactical and strategic decisions. Download our guide below to learn more about the skills you’ll need to accelerate your career and how an advanced degree can help.