Basics

Data visualization - What is it and why is it important?

Tables and figures make your data analysis easy to understand, making data visualization an essential part of any research project. Learn how data visualization can help you make your research more impactful.
Roehl Sybing
Content creator and qualitative data expert
  1. Data visualization - an Overview
  2. Data visualizations
  3. From raw data to data visualization
  4. Data visualization software
  5. Data visualization of big data
  6. ATLAS.ti's data visualization tools
  7. Which is the right data visualization tool for your research?

Data visualization - Overview

How often do you read a research article and skip straight to the tables and figures? That's because data visualizations have the power to make large and complex research projects with thousands of data points comprehensible when presenting data to research audiences. Good data visualization tools can help you summarize your data and make clear the pathways for actionable insights.

Figure 1: Line charts help illustrate the research in a report. Photo by Markus Spiske.

Let's take a look at the process researchers employ to create data visualizations and how ATLAS.ti can assist you with providing a visual representation of your research.

  • Data visualizations
  • From raw data to data visualization
  • Data visualization software
  • Data visualization of big data
  • ATLAS.ti's data visualization tools
  • Which is the right data visualization tool for your research?

Data visualizations

In everyday situations, a picture is always worth a thousand words. Illustrations, figures, and charts convey messages that words alone cannot. In research, data visualization can help explain scientific knowledge in an orderly manner based on data that is otherwise unstructured.

For all of the various data formats available to researchers, a significant portion of qualitative and social science research is still text-based. Essays, reports, and research articles still rely on writing practices aimed at repackaging research in prose form. This can create the impression that simply writing more will persuade research audiences. On the contrary, framing research in terms that is easy for your target readers to understand makes it easier for your research to become published in peer-reviewed scholarly journals or find engagement at scholarly conferences. Even in market or professional settings, data visualization is an essential concept when you need to convince others about the insights of your research and the recommendations you make based on the data.

What is challenging about the written word is that the ability to make data-driven decisions relies on a basic understanding of the data collected during research. However, complex data sources can produce big data, making it challenging to summarize or reduce large amounts of information to simple verbal descriptions.

Data visualization can address this challenge. Conceptually, employing data visualization techniques in research writing and presentation aims to provide a means for understanding otherwise unordered or unstructured data points. They can represent the data and your analysis in a concise manner where words alone might not be the most efficient way to convey your message.

Data visualization is important because it makes it easy for your research audience to understand your data sets and your findings. Also, data visualization helps you organize your data more efficiently. As the explanation of ATLAS.ti's tools will illustrate later on, data visualization might point you to research inquiries that you might not even be aware of, helping you get the most out of your data.

Figure 2: A bar chart is a useful data visualization. Photo by Morgan Housel.

Research employs numerous visualizations to help researchers explain their data, including:

  • tables
  • line charts
  • bar charts
  • pie charts
  • scatter plots
  • pictographs
  • word clouds

Whatever the data visualization, a good visualization succinctly conveys the main points of the research findings while also illustrating the data points in such detail to allow research audiences to understand the data and the analysis.

Ultimately, the task of visualizing data is unavoidable in data analytics contexts where making sense of big data is impossible to frame through words alone. Even in social sciences and other scholarly disciplines, publishing a research manuscript depends on explaining to reviewers how you collect and analyze data, making data visualization important.

For powerful data visualization, start with the right research tool.

Rich data visualizations start with meaningful analysis with ATLAS.ti. Start with a free trial.

From raw data to data visualization

Strictly speaking, the primary role of data visualization is to make the analysis of your data, if not the data itself, clear. Especially in social science research, a data visualization makes it easy to see how data scientists collect and analyze data.

Think about a minute-by-minute description of a soccer match, for example. The sequence of events describing one play after another, who has the ball, who passes it to whom, who scores a goal, and how they scored it make up the raw data to be analyzed. Pundits analyze the data by generating statistics such as goals, assists, key passes, possession ratios, and goalkeeper saves. These statistics are visualized in all sorts of ways in box scores and player profiles, which help fans understand the aspects of a particular game or player and make determinations about the quality of the game they are watching.

Figure 3: Raw data without data visualization can be hard to understand. Photo by Wesley Tingey.

In the above example, the visualizations depend on the statistical analyses that pull the relevant information from the raw data. As a result, to rigorously visualize data, you need to have a well-intentioned strategy for analyzing raw data. In other words, how do you look at the data? What are you looking for when you examine your data?

The decisions you make to collect and analyze data will inform your data visualization. For example, if you are conducting market research on customer satisfaction of a product, you might think about the following questions:

  • How do you measure customer satisfaction?
  • What is the definition of a "satisfied" customer?
  • What data should you look at to measure customer satisfaction?

Perhaps you might survey customers and ask them to rate the product on a ten-point scale, in which case a simple bar graph might show a distribution of customers along that scale to get a sense of where the market stands on your product. On the other hand, you might also interview customers and listen to what positive and negative keywords they use to talk about a product. As a result, a word cloud might be the most appropriate data visualization for that analysis. Whatever analysis you pursue, choosing the best data visualization for your research is an important decision to make.

From small projects to big data, visualize it with ATLAS.ti.

A free trial of ATLAS.ti will show you what you can do with your data.

Data visualization software

Of course, data scientists with a comprehensive understanding of their data can manually produce meaningful data visualization in platforms like Microsoft Excel and Google Drawings, but these are self-service programs that aren't integrated with analytical tools like in ATLAS.ti. Contemporary qualitative researchers turn to programs like ATLAS.ti to perform the analysis of their data and produce the corresponding visualizations.

First, however, it is the organization and coding of research that are essential to visualizing your data. As intuitive as the best data visualization tool might be, the preparation of collected data is essential to make the analysis and visualization processes easier and quicker, both for you and other users.

Organizing data

To facilitate data visualization, it is important to consider how the data in your research project should be organized. A market survey might produce hundreds or thousands of response records, so you might want to separate the responses of satisfied customers from the responses of unsatisfied customers to analyze the differences. Likewise, if you are conducting an interview study, you might benefit from separating documents by interview respondent characteristics.

However, what if you have conducted multiple interviews with respondents? Document groups in ATLAS.ti can help you categorize documents that share a common characteristic. Document groups provide a convenient unit of analysis for many of the analysis tools in ATLAS.ti. You can think of a document group like a folder that holds a set of individual documents.

Figure 4: Data sources can be grouped together in document groups in ATLAS.ti. Photo by Viktor Talashuk.

Unlike physical documents in file folders, documents in ATLAS.ti can be assigned to multiple groups which can prove useful for later data visualization. In the example of the interview study, documents representing individual interviews can be categorized into multiple groups along lines of gender, age group, ethnicity, or customer satisfaction.

In conjunction with coding and data visualization tools in ATLAS.ti, document groups can help you create useful visualizations comparing broader sets of data. Instead of comparing individual interview respondents, for example, you can compare customer satisfaction across different age groups or ethnicities.

Coding data

Data visualization is effective in explaining research to others only if the researcher or data scientist can make sense of the data in front of them. Traditional research with unstructured data calls for coding the data with short, descriptive codes that can be analyzed later, whether statistically or thematically.

Figure 5: In traditional coding, highlights and flags help to visualize data to researchers. Photo by Russ Ward.

These codes form the basic data points of a meaningful qualitative analysis. They represent the structure of qualitative data sets, without which a scientific visualization with research rigor cannot be achieved. In most respects, data visualization of a qualitative research project requires coding the entire data set so that the codes adequately represent the collected data.

Data visualization of big data

Big data, however, poses a significant challenge for this process, especially when the enormous amount of work involved with analyzing big data makes a manual coding process leading up to data visualization all but infeasible. As a result, data science has developed artificial intelligence and machine learning to analyze and predict the meaning of complex data.

ATLAS.ti has a number of AI tools that assist with the coding process so that you can create visualizations of big data, business data, or simply data from projects of any size or scale.

AI Coding

By applying machine learning to big data, ATLAS.ti can automatically suggest codes for your data, saving time when working on large projects and making smaller projects easier when you are looking for useful insights or new avenues to analyze data.

Figure 6: AI Coding is one of the top data visualization tools in ATLAS.ti.

AI Coding automatically gives you a report on the codes it most frequently applies to your data, as well as which codes co-occur together most often. This is a useful tool for data visualization of concepts that artificial intelligence detects in your research project.

Named Entity Recognition

AI can recognize proper nouns (e.g., New York, Oxford University) from common nouns (e.g., tree, book), which is a useful capability when you are conducting research on brands, companies, institutions, or groups of people. ATLAS.ti packages this AI capability in its Named Entity Recognition tool to make it easier for data scientists to apply codes to proper names.

Figure 7: The Named Entity Recognition tool in ATLAS.ti helps researchers analyze data.

Suppose that you are conducting interviews for market research where respondents are asking about the most common brand names they recognize. If the objective of your research is to determine the most recognizable brands, Named Entity Recognition can examine interview transcripts, identify those brands, and apply corresponding codes to those data segments. You can then look at the frequencies of codes in your project to get a sense of which codes, and thus which brands, are mentioned most often by interview respondents.

Opinion Mining

This powerful AI tool is a combination of ATLAS.ti's Sentiment Analysis and Concepts tools. Sentiment codes can give a sense of whether a segment of text is positive or negative, while a concept in ATLAS.ti is a combination of words that form a particular meaning (e.g., "nature sounds" versus "sound mind and body"). You can also examine visual concept clouds in ATLAS.ti's Concepts tool. In Opinion Mining, you can then take your analysis further by coding these two aspects in tandem to produce deeper insights than can be derived from one aspect alone.

ATLAS.ti's data visualization tools

Among the various data visualization tools on the market, ATLAS.ti is an ideal data analytics platform for producing visual representations of qualitative data. Whether you are developing business intelligence or social science theory, ATLAS.ti has multiple data visualization methods to help illustrate the analysis of your research to yourself and your research audience.

Code Manager

The Code Manager is the central place where all of your codes are stored and organized. In the Code Manager, each code has groundedness and density numbers, showing how many times a code has been applied and how many codes are connected to that code, respectively. You can use these statistics in a data visualization to highlight the theoretical development of a code in your project.

Figure 8: The Code Manager in ATLAS.ti provides a visual representation of codes.

Moreover, the interactive data visualizations in Code Manager can help you navigate between your codes and their associated quotations when you just need to sift through specific data sets. You can comfortably view code distribution bars and code frequency clouds, bars, and TreeMaps.

Word Frequencies

In Word Frequencies, the word cloud is a useful data visualization tool for showing which words appear more frequently than others. You can narrow the analysis to a set of documents or a set of quotations from a particular code. Then you can even adjust the settings to narrow the word cloud to certain parts of speech. The resulting word cloud illustrates which words are used most often in your data.

Figure 9: The Word Frequencies tool in ATLAS.ti has various visualization methods.

Co-Occurrence Analysis

Co-Occurrence Analysis is a powerful tool to demonstrate the relationship between codes, represented by the number of times they "co-occur" or are applied together in your data. In qualitative research, this co-occurrence is essential to understanding the extent that different social phenomena correlate with each other.

For example, suppose that your research inquiry relates to sentiments about brand names. If you have used the Named Entity Recognition to identify brands in conjunction with the Sentiment Analysis or Opinion Mining tools to code for positive and negative sentiments, you can cross-tabulate codes of brand names with codes of sentiments to determine whether brands co-occur with positive or negative sentiments.

Data visualization in the Co-Occurrence Analysis tool can take the form of a simple table, a bar chart, or a Sankey diagram. The Sankey diagram is a powerful data visualization that illustrates the strength of a relationship between codes relative to other relationships.

Figure 10: The Sankey diagram in ATLAS.ti is a powerful graphical representation of the relationships in your research.

Code-Document Table

The Code-Document Table works in a similar fashion to Co-Occurrence Analysis in that you can perform data visualization through charts, tables, or Sankey diagrams. The Code-Document Table is useful for illustrating relationships between documents and codes.

Figure 11: The Code-Document Table tool in ATLAS.ti helps you understand the application of codes in your data sets.

If you have structured your data by documents (e.g., interviews with different respondents, observations by day, etc.), a Code-Document Table can give you a sense of which codes are associated with which documents. With this data visualization, you can distinguish between different types of documents (e.g., interview respondents with positive feelings, interview respondents with negative feelings) and narrow your analysis accordingly.

Networks

When you need a data visualization that illustrates theories or concepts generated from your data, you can use the Networks tool to draw hierarchical or otherwise relational connections between your codes. A network is a series of nodes that represent the various elements of your project (e.g., codes, quotations, documents, memos, groups). When visually connected through links, a network can provide a visual sense of the theory or framework that you derive from your research.

Figure 12: The Networks tool in ATLAS.ti provides a scientific visualization of the various elements of your project.

TreeMaps

The TreeMaps tool produces an illustrative data visualization of codes and documents, where the development of each project element is represented by the space it occupies in the TreeMap. Just like with word clouds and words, a TreeMap of a project's codes can tell you which codes appear most often in your data. Their prominence in a TreeMap can indicate where the focus of your research lies, and clicking on any space in a TreeMap will display the relevant quotations in one place to allow for easy analysis of the most meaningful data.

Figure 13: Create visualizations with the TreeMap view in Code Manager in ATLAS.ti.

Other visualization tools

At certain points in the data analysis process, a more detailed, graphical representation of your data will require tools outside of ATLAS.ti. Researchers often use a data analysis program such as ATLAS.ti in tandem with programs like Microsoft Excel when a deeper, quantitative analysis is necessary. You can import data from ATLAS.ti to Microsoft Excel to create data visualizations such as scatter plots, pie charts, and line graphs.

More advanced Microsoft Excel users can also create interactive charts, but it all depends on collecting and analyzing data from data sources through data analysis software like ATLAS.ti.

Which is the right data visualization tool for your research?

You can have the top data visualization tools from ATLAS.ti to make your research clearer to your audiences, but how do you choose the best data visualization tool for your project? As with any analytical tool, the right visualization tool depends on the nature of your research inquiry.

Describing concepts

One main purpose of qualitative research is to describe the aspects of a social phenomenon. What factors do people consider important to defining happiness? What qualities make a good movie? What are the different considerations regarding customer satisfaction?

In ATLAS.ti, you can use data visualization tools to illustrate the most frequently used codes in your project. TreeMaps can help identify qualities or considerations represented by your project's codes and illustrate it to your research audience.

Drawing connections

Alternatively, you can take the most frequently occurring codes and represent and link them in a network to further illustrate correlations and causal relationships within a working theory represented by codes. A visualization of a theory identifies key elements of social phenomena in unordered, raw data as well as the connections between them.

The Co-Occurrence Analysis and Code-Document tools can create visualizations like bar charts or Sankey diagrams to provide supporting evidence of the connections in your data. Networks are also useful for understanding connections between the elements of your project. Within networks, you can manually establish links between codes if you perceive a relationship between codes that you want to explain to your research audience.

To tell the story of your research and convey your impacting findings to diverse audiences with various visualizations, you can count on ATLAS.ti.

Find the best data visualization tools in ATLAS.ti.

Start with a free trial of our powerful data analysis platform today.