Often data visualization is interpreted as just visualizing data effectively. However,  it is much more than just creating visually-appealing graphs/charts. Before giving your data a voice, you need to understand the different data types you'll be dealing with in order to decide the best chart type for your data.

In this blog post, you'll learn about the different types of data and their measurement scales.

Data Types

1. Qualitative Data - Any information that is non-numerical in nature is termed as Qualitative Data. This type of data is categorized based on attributes, properties, and labels and describes information in the form of open-ended answers allowing the respondents to express themselves.

2. Quantitative Data - Any information whose value is measured in the form of numbers is termed as Quantitative Data. This data type can likewise be characterized as a group of quantifiable information that can be used for numerical calculations and statistical analysis, influencing real-life decisions.

Measurement Scales

In Statistics, the numbers or variables are defined and categorized using different measurement scales. Each measurement scale has a specific property that helps to interpret the numbers assigned to people, phenomenon, objects, and events.

There are 4 measurement scales:

1. Nominal Scale - Data Type : Qualitative

Nominal Scale is a measurement scale where numbers are served as "labels" to identify and classify an object. This type of scale does not require the use of numerical values or order.

Examples of Nominal Scale -
a. Gender : Male, Female
b. Eye Color : Black, Brown, Blue
c. Pet : Cat, Dog, Fish

Characteristic of Nominal Scale -
Nominal data can be quantitative as well. For example, zip code. However, the quantitative labels lack a relationship and mathematical operations cannot be performed on them.

How to analyze nominal data?
To analyze nominal data, you can use the grouping method. The nominal variables are grouped and classified into various categories, and for each category, the frequency (mode) or percentage can be calculated. The bonus point is you can apply statistical methods to analyze nominal data. For example, you can analyze this type of data using hypothesis testing i.e. using Chi-Square test. After conducting the analysis, nominal data can be graphically represented using bar chart or pie chart.

2. Ordinal Scale - Data Type : Qualitative

Ordinal Scale is a measurement scale where the order of the variables matter but not the difference between each of these variables.

Examples of Ordinal Scale -
a. Socio-Economic Background : Upper Middle Class, Lower Middle Class
b. Class Ranking : 1st, 2nd, 3rd
c. Frequency of Occurrence : Always, Very Often, Sometimes, Rarely, Never

Characteristics of Ordinal Scale -
a. Ordinal Scale helps surveyors analyze the degree of agreement/disagreement among respondents with respect to the identified order of the variables.
b. This type of scale has no zero point i.e. it is either arbitrary or absolute.

How to analyze ordinal data?
The simplest way to analyze ordinal data is to present it in tabular format in which each row indicates a distinct category. However, to compare two ordinal groups, you should use the Mann-Whitney U test. This test helps us to analyze if a variable from one sample is greater or lesser than another variable randomly selected from a different sample.

3. Interval Scale - Data Type : Quantitative

Interval Scales are numeric scales in which there is order and the difference between two values is equal and meaningful.

Examples of Interval Scale -
a. Temperature : Fahrenheit, Celsius
b. Credit Score : 300-850
c. Measuring income as a range : $0-$999, $1000-$1999, etc.

Characteristic of Interval Scale -
Arithmetic operations like addition and subtraction can be performed on interval data. However, mathematical operations like multiplication or division cannot be applied on interval data.

How to analyze interval data?
Since interval data is quantitative in nature, almost all statistical analysis can be applied when calculating interval data. Some of them are mean, median, mode, standard deviation, trend analysis, conjoint analysis, SWOT (Strengths, Weaknesses, Opportunities and Threats) analysis, and TURF (Total Unduplicated Reach and Frequency) analysis.

4. Ratio Scale - Data Type : Quantitative

Ratio Scales are created by units of equal size, same as an interval scale with the added condition that "zero" is not arbitrary. In simple words, it is an interval scale with a true "zero".

Examples of Ratio Scale -
a. Age : Less than 20, 21-30, 31-40, etc.
b. Weight in Pounds : Less than 60, 61-120, 121-160, etc.
c. Height : Less than 5 feet, 5 feet 1 inch - 5 feet 5 inches, etc.

Characteristics of Ratio Scale -
a. Ratio Scale can be measured, ranked, and ordered.
b. Data can be added, subtracted, multiplied or divided.

How to analyze ratio data?
Any type of statistical tests i.e. either parametric or non-parametric test can be performed on ratio data since it allows all mathematical operations. However, parametric tests are more powerful because it allows you to make stronger conclusions.

I hope this blog post helped you discover the different data types and the statistical measurement scale you can use at each data type.

1. Do you want to learn about data visualization's history, theory, principles, and more?
2. Do you want to create effective visualizations that tell a story?
3. Do you want inspiration from stunning visualizations?

Here are a couple of books that I have found invaluable and highly recommend to anyone who wants to gain a deeper understanding of data visualization.

1. Storytelling with Data : Cole Nussbaumer Knaflic

If you're new to the data visualization world, then this book is a must read as it covers the fundamentals of data visualization. The structure of this book is oriented towards pragmatic data visualization where the author uses real-world examples to explain how to effectively communicate data to your audience. These examples also show ordinary graphs/charts can be enhanced by reducing clutter, using better color combinations, and other standard tricks.

The USP for me is the way Nussbaumer Knaflic introduces a first version of a dataviz, then walk through her thought process of eliminating clutter and emphasizing on just the key information, thus telling a story. She finishes the book by focusing on the fact that it is important to iterate and modify designs and test them on various audiences to see how they interact with the visualization.

In summary, I definitely recommend "Storytelling with Data" as it gives you a different perspective on what data to present and how to build a story across that data.

2. The Visual Display of Quantitative Information : Edward R. Tufte

The theme of this book is to be able to represent complex data using simple graphics. Tufte also provides the history and evolution of various graphs such as time-series, maps, and narratives or stories. This book is basically about graphical representation of data where Tufte introduces theories, terminologies, and how to communicate complex ideas with clarity and precision to maintain graphical integrity and excellence. It explains the reason for poor quality of data graphics : where one transfers the design ownership to those with formal art/design training but without an understanding of statistical methods to analyze the data.

The author also provides suggestions and examples on how to design box plots, bar graphs, and scatterplots. What I loved the most is the author's idea of a "range-frame" where you draw the x and y axis on a scatterplot, but only in the range where the data is present.

If developing data graphics is a part of your routine, then this is an essential addition to you book collection. Because after reading, you will likely look at data graphics with an eagle eye and learn to establish a structured process for creating efficient data graphics.

3. The Functional Art : Alberto Cairo

This book is a perfect blend of statistics, design, and data journalism that weaves visualization theory and techniques with real-world stories and interviews about existing visualization projects. As you flip though, you'll notice a ton of examples and pencil sketches with an explanation of the process.

The first section of the book lays the "foundation" and explains the benefits and techniques to a good visualization. Important points are displayed in bold text which makes it easier to spot.

The second section covers "cognition" i.e. it explains how human brain perceives different visuals in the form of color, shape, size, etc.

The third one is a "practice" section where  the author provides diverse case studies and interactive graphics with the process behind it. I love when authors explain their process behind a beautiful end result as it helps you understand the effort it takes to create a stunning visual.

The last section "profiles" is a collection of interviews with some of the luminaries of the data visualization space: Hans Rosling, Jan Willem Tulp, Hannah Fairfield, Geoff McGhee, and Juan Velasco, to name a few.

In summary, this book does a fantastic job of clarifying the dynamics behind visual design. Something that I acknowledged the most about the author is that, he brings academic visualization concepts into his book with great examples, making it truly valuable to a wide range of audience.

4. The Truthful Art : Alberto Cairo

This book comes in handy when you want to visualize data in a way where anyone can understand and make data-driven decisions. The author starts with a basic introduction to statistics and visualization. It includes how to utilize different charts and graphs to explore and extract insights from your data. Every now and then statistical concepts such as confidence intervals, error margins, and hypothesis testing are explained in a simple language. But some chapters dig deeper into certain concepts of statistics, like correlation, regression, distributions, statistical significance, etc. The book is also loaded with beautiful examples of maps and tips on how to avoid common data pitfalls.

Most of these concepts and techniques were already introduced to me while pursuing my Masters. However, I still found it to be valuable as it helped me apply some of these concepts while creating data viz projects.

If you're a data journalist or data visualization designer, then this book is a must read as it helps you understand and lay the foundation towards a good visualization. It will also help you understand how data viz experts think while creating stunning visuals.

5. The Big Book of Dashboards : Steve Wexler, Jeffrey Shaffer, and Andy Cotgreave

When you want to build a dashboard to summarize your findings, where do you start? Do you just combine a set of charts or scorecards and call it a dashboard, without even knowing the techniques or best practices? If your answer is "Yes", then you should grab a copy of this book. Because this book starts with practical design advice and then moves to 28 real-world case studies with best dashboard designs in practice.

The early chapters cover data types, encoding data, color coding, preattentive attributes, and common chart types. This acts as a foundation to understand the case studies.

The latter chapters are filled with dashboards addressing different industries and platforms (desktop, tablet, smartphone, and print) with a detailed analysis of which visual/chart "works" and which "does not". Further, the case studies discuss the common business situations encountered by analysts and this section is the meat of this book. The authors also discuss about the iterations some of the dashboards went through and why the final forms were more successful.

Overall, if your focus is to overcome business problems through design, regardless of the tools you use, then this book should be on your reading list.

6. Resonate : Nancy Duarte

The theme of this book is "persuasion." It explains how to understand your audiences', create persuasive visuals, and build a presentation to grab their attention. Duarte emphasizes that in order to develop a deep connection with your audience, it is important to have an emotional allure to introductions and presentations, not just presenting statistics.

Using the fundamental concept of storytelling that resonates with your audiences' is one of the most incredible methods to evoke emotional responses. You create desire and build trust and credibility through stories that drives people to adopt your perspective. The author presents her ideas through graphs, charts, illustrations, photographs, and case studies so that the reader understands the broader meaning of her message.

However, the book utilizes audience analysis heavily only in the early chapters i.e. until chapter 4.  In the later chapters the author discusses on how to select, organize, and structure meaningful content. I think there's an opportunity to highlight that audiences' requirements should drive the selection, organization, and structure of the content. Apart from that, I think this book serves as a brilliant guide to anyone who communicates and persuades on a daily basis.

7. Understanding the World : Sandra Rendgen

This is a beautiful giant book that utilizes artistic design to convey information that everyone can understand. This book is divided into 5 chapters - environment, technology, economics, society, and culture to explain complex problems in the form of infographics. You'll love and enjoy turning each and every page of this book because every piece of infographic is stunning, unique, and loaded with information. This book is a great resource if you want some inspiration while trying to visualize your own data.

Let's begin with understanding what data visualization is all about.

In simple words, data visualization is the art of transforming quantitative data into visuals that are easier for the human brain to grasp and understand.

Believe it or not - Data visualization is the new storytelling since it is one of the most powerful modes to communicate data through visuals. When done right, it delivers information which leads to good decision-making, but when done wrong, it can lead to muddled conclusions which in turn can hamper your business.

The goal of this blog post is to give you a clear idea of the concepts that make a good visualization.

1. Start off by asking the fundamental question - Do I have the required data?
You might have heard how valuable data can be - to improve business processes and to make better decisions. However, there's an admonition - data is valuable only if it's of high quality. So, how do you measure your data quality?
There are 5 main criteria to measure your data quality:
a. Accuracy - This is one of the most important aspects of data quality. For data to be accurate, all the values must be represented in a consistent and unambiguous form. This prevents confusion and guarantees there is no ambiguity when analyzed by a computer.
b. Relevancy - Data relevancy is the degree to which the information being gathered meets the purpose for which it is gathered. It should indisputably report all the positive and negative information. Only then, it can act as the ace of spades.
c. Completeness - Incomplete data is as dangerous as inaccurate and irrelevant data. Gaps/missing fields in data collection leads to partial view of the overall picture. Therefore, data owners must ensure there is no data missing from the dataset. Setting up automation to check data gaps and duplicates will help you avoid time wastage.
d. Timeliness - Timing is everything when you're collecting, analyzing, and visualizing data. If a business looks into outdated data, even if only for a few weeks or months, it can make a huge difference. Hence, having up-to-date data is as important as having accurate data.

2. There is no need to explain the significance of data to yourself when you're the data owner. However, for the larger audience you must provide a clear explanation of what the data is about, and that’s where a good design comes into the picture.

"Information graphics should be aesthetically pleasing but many designers think about aesthetics before they think about structure, about the information itself, about the story the graphic should tell" - Alberto Cairo

Compelling visuals should involve the following design principles:
a. Research - When starting a project and trying to find out what the solution could be, we need to conduct research to get the inspiration for an answer. The goal of design research is to help us understand the limits of various standards and guidelines for creating powerful visualizations. For that, start asking the right questions -

• What information do you want to communicate?
• Who are our audiences?
• What is their level of numeracy and visual literacy?
• How much time your audiences can spend?
• How much precision is required?

If you stop and think about these questions, this is pretty ambitious target! And once you've answered these questions, start planning and designing solutions in a way that aligns the interests of the audience. After all, design research is all about covering a wide variety of topics from observation to memorability, from simple/complex frameworks to hypothesis about what comprises a chart.
b. Strategy - Building a design strategy requires you to create the corporate strategy around design thinking and human-centered design approach. Start by developing a framework towards how you want to organize the data, so that, it is understandable, techniques to create enticing visuals, systems to store the data, and finally the tools to deliver them.
Our eyes are naturally attuned to search for things that stand out. And with a design strategy in place, you can influence how your data will be processed and experienced.
c. Exploration - Now that we've a got a clear idea of what research needs to be conducted and what strategy needs to be developed, it's time to conceptualize ideas and promote creativity and innovation within yourself. This is where tools like GitMind can be used to create wireframes. Once you've got these wireframes in place, start working on visual cues by choosing apt color, form (lines, dots, and shapes), depth (space, size color, lighting, and textural gradients), and movement.

3. Lastly, a good user experience depends on usability i.e. how well the end user can use your visualization to achieve a defined goal effectively.
Now that you have the required data, research, strategy, and design exploration in place, it's time to create an impactful visual. Break down complex pieces of information into visual representation that is cleaner and easier to understand. Format information into digestible bits - add annotations, so that, the end user can easily identify and pick out the most useful pieces of information.

Most businesses need to derive insights from huge amounts of data on a daily basis. They spend hours to crunch data and produce reports or briefings just to ensure that they can be digested by their clients and colleagues. We all might agree that crunching data is important. However, it is not enough - you need to be telling stories in the form of visuals. A good visualization saves time and makes it easier to identify trends and patterns which otherwise is difficult when you're reading a report with hundreds of pages.