All businesses need to perform data visualization to make any sense of the information available to them. Increasingly, this is becoming synonymous with online digital visualization as IT networks take up an ever more central information processing role in companies.
This article looks at what data visualization actually is, how scientific and artistic processes are both important when creating effective ways to visualize data and why this element of data processing will become increasingly important in the coming years and decades.
Data visualization in a nutshell
Without the benefit of data visualization aids, understanding what is going on within a business is reduced to guesswork. As data visualization becomes more intuitive and instructive, the easier it becomes to draw meaningful conclusions from the raw numbers in order to inform business strategy or deliver a snapshot of current performance.
Data visualization is also central to running business admin operations, call handling, accounting, auditing, cyber security, application development, assessing compliance, HR, IT support, training and many other areas of the business. It can be used to identify patterns, highlight omissions and compare current performance against a company or industry benchmark.
Data visualization needs might be broad, focusing on information which is relevant to the company as a whole, or narrow, honing in on individual departments, specific systems or even personnel within those departments.
There are many different ways in which data can be organized and presented. Some of these will succeed in revealing the information hidden in the numbers while others will obscure or even distort that information.
Therefore, it is critical that the correct methods of data visualization are selected according to business need.
Before looking in depth at what options are available, it is useful to look at the purpose of collecting data in the first place. What happens (or should happen) to the raw attributes and variables collected and measured in the field? The data processing cycle below, adapted from one created by Cathy O’Neil and Rachel Schutt (2014) in ‘Doing Data Science,’ is fit for that purpose.
The data processing cycle
The data processing cycle can be thought of as the science behind data visualization. It is a method which can be applied to all data being collected by a business, helping to rationalize and systematize its treatment.
All data originates as a measurable fact or process. The temperature of a refrigerator, the movement of funds from one bank to another, the submission of an email address via a form field, the clicking of an online ad are just a few examples of the facts and processes that could be recorded and abstracted into data.
The raw data collected then has to be processed and converted into an organized dataset (often a table of some kind). This dataset is then fed into a model or algorithm (with or without initial exploratory analysis) before being presented as some form of digital or physical graphic or report.
In some systems, data might also be entered into a ‘data product,’ an application designed to use the data it receives to create more useful data â€“ and so the cycle continues. An oft-quoted example of an effective data product is Amazon’s ‘recommendations’ program which takes the output of an algorithm to stimulate more sales.
However, where humans are concerned, we need the intermediate step of data visualization in order to make decisions about whether and how to change the way a person, system, department or the business as a whole carries out their day-to-day activities.
The role of form and aesthetics in data visualization
According to statistician Edward Tufte, good data visualization presentations should achieve a number of things, including:
- Encourage viewers to think about the methodology behind the data
- Avoid distortion of the data
- Facilitate several layers of detail
- Integrate well with statistical and verbal descriptions of the data
Whether data visualization is viewed offline as a printed report, on a computer program (e.g. a Microsoft Excel table) or as part of a customized real-time admin dashboard, it is important that the correct form of presentation is chosen. For example:
- Tables for looking up specific values
- Bar graphs for spotting patterns
- Flowcharts for understanding systems
- Pie charts for visualizing proportions
- Scatter graphs for assessing correlations
- Histograms for recording frequencies
- Cartograms for presenting geographically-relevant detail
All information necessary for interpreting the data (e.g. correctly labeled axes, chart keys, units, etc.) should be available with the most important information readily to hand.
The art of data visualization focuses on communicating the relevant information in an engaging way through the use of color, proportion and spatial arrangement. These are termed ‘pre-attentive attributes’ as they initiate the visualization process before the viewer has even begun to consciously attend to the graph, chart, table or other graphic.
Some data visualization applications will allow fine customization of individual dashboards so that displays can be tailored on a departmental or even individual level. It is worth trialling several different products (or in-house designs) before committing funds. A good IT consulting firm may also be able to help with sourcing discounted solutions.
In order to balance the functional and aesthetic purposes of data visualization, designers need to be mindful of the presence of ‘chartjunk’ which is anything that doesn’t add value to the presentation.
Why data visualization will matter more than ever in the IoT age
The advance of technology has made it easier to capture, store and process data than most people could ever have imagined. The age of ‘Big Data’ promises increasing granularity as almost every element of every business system is squeezed for information. Those businesses that are able to use that rich data for making smart business decisions will rise to the top of their industry. That calls for optimized data visualization tools and processes.
By incorporating deep learning and other AI technologies, data products are also going to become much more sophisticated. What’s more, they are going to be increasingly interconnected as part of the Internet of Things. Data visualization will need to play its part as a ‘glass pane’ through which humans can understand what is going on beneath the hood. Only then will our businesses become the beneficiaries of the technology we’re creating.
Brent Whitfield is the CEO of DCG Technical Solutions Inc. DCG provides the specialist advice and IT Support Los Angeles area businesses need to remain competitive and productive, while being sensitive to limited IT budgets. Brent has been featured in Fast Company, CNBC, Network Computing, Reuters, and Yahoo Business. https://www.dcgla.com was recognized among the Top 10 Fastest Growing MSPs in North America by MSP mentor. Twitter: @DCGCloud