Discovery through Visualization
By Chris Long, Customer Intelligence Solutions Engineer, SAS Canada
Data visualization is the presentation of data in a pictorial or graphical format. Though the term may be new, the concept is not. For centuries, people have depended on visual representations, such as charts, graphs, and maps to understand information easily and quickly. As more and more data is collected and analyzed, decision makers at all levels will increasingly look to data visualization software to find relevance among millions of variables, communicate concepts and hypotheses to others, and even predict the future. Today’s organizations need to think of their data as a great but unedited story which needs the help of visual analytics to bring it to life.
In a big data world we need data visualization:
According to Research Scientist Andrew McAfee and Professor Erik Brynjolfsson of MIT, the amount of data that crosses the Internet every second is greater than all the data stored in the Internet just 20 years ago. This amounts to exabytes of data being created on a daily basis. While big data remains, for some, an overhyped term, the reality is that the explosion of data is unlike anything we’ve seen before. Data volumes will continue to grow as more devices come online including computers, smart phones, tablets, new apps and services, along with an increasing number of devices outfitted with smart meters and sensors and GIS transmitters.
In light of this data explosion, data visualization tools are becoming increasingly important for managing and navigating information glut, and are making big data easier to digest. With advancements in technology, data visualizations are taking on more complex forms than ever before. They are being used to unravel the meaning behind big data sets that would otherwise be too difficult to understand.
How data visualization is changing the world of marketing:
Most marketing organizations are awash with information, whether it’s demographic, socioeconomic, geographic, behavioural, transactional, or one of the other myriad of data segments that make up customer intelligence like social media and smart devices. However, this information is often stored in a variety of different systems, some internally owned by different—often siloed—departments and articulated in various formats but also external in repositories like Datasift and GNIP. What is lacking from this information surplus are the insights needed to make the best marketing decisions. Presenting information visually empowers marketers to very quickly explore all customer data, no matter the size. Visual data adds a level of accessibility that can help marketers rapidly identify key relationships and uncover insights for creating more detailed customer segments (e.g., based on purchase history, sentiment towards brand, life stage, etc.) and more personalized promotions and messages.
Today’s marketers need to be nimble at converting data into insight, and data visualization software is integral to helping them find relevance among the millions of variables that can help target customers with relevant offers. For marketers, that can mean decreasing the time spent on finding key answers such as which messages are best suited for different customers, which time of day or day of the week are more important for certain offers, or figuring impact of activities to better plan marketing resources.
While data visualization is a very glamorous field right now, users must be pragmatic and start with the basics: data analysis, data management and best practices in data visualization. If possible, try before you buy the solution and make sure you are asking questions that are realistic and tied to your business goals. Most importantly, ask yourself, will it help you forecast? A lot of tools enable hindsight but does the tool offer foresight – can it offer you actionable benefits to plan for future success?
In order to generate the best visualization for your data there are a few basic concepts you should follow:
- Understand the data you are trying to visualize, including its size and cardinality (the uniqueness of data values in a column)
- Wherever possible remove noise and any data deemed obviously irrelevant.
- Determine what you are trying to visualize and what kind of information you want to communicate
- Know your audience and understand how they process visual information
- Use a visual that conveys the information in the best and simplest form for your audience
A picture is worth a thousand words: What is your data telling you?
Data visualization is an art and a science unto itself, and there are many graphical techniques that can be used to help people understand the story their data is telling. Here is a quick look at some chart types used for visualizations of data.
Line Graphs: Shows the relationship of one variable to another. They’re most often used to track changes or trends over time (see Figure 1).
Bar Charts: Most commonly used for comparing the quantities of different categories or groups.
Scatter Plot (or X-Y plot): A two-dimensional plot that shows the joint variation of two data items. In a scatter plot, each marker (symbols such as dots, squares and plus signs) represents an observation. A scatter plot is a good way to visualize relationships in data.
Pie Charts: Pie charts are most effective when there are limited components and when text and percentages are included to describe content (see figure 4).
Word or Phrase Cloud: Data variety brings challenges because semi structured and unstructured data require new visualization techniques. A word or phrase cloud visual (where the size of the word or phrase represents its frequency within a body of text) can be used on unstructured data as a way to display high or low frequency words or phrases (see Figure 5).
Network diagram: Network diagram explore relationships within a data set, including connections across geographies (see figure 6).
Enabling the “Data Scientist”:
As more organizations embrace data as a strategic asset, we will see a rise in the need for individuals who are capable of extracting value from these data assets. Some call them data scientists others call them data engineers – regardless of their title, these individuals possess the ability to extract value from data, the kind of insight that leads to sound business decisions and increased ROI. McKinsey predicts a shortage in the coming years of big data scientists and big data managers.
The appeal of data visualization tools is that the need for “data scientists” are reduced because organizations can now put these kind of tools in the hands of a business user. Visualization based solutions offer ease of use and enable users to explore the data, find patterns and spot inconsistencies without much training. With visualization tools users don’t need to write code or understand modeling – the tools do the heavy lifting so that the end user can focus on what story the data is telling them.
Because of the way the human brain processes information, it is faster for people to grasp the meaning of many data points when they are displayed in charts and graphs rather than poring over piles of spreadsheets or reading pages and pages of reports. A picture is worth a thousand words really is necessary to cut through the crowded, big data-cluttered world we currently live in.
Originally published in VUE Magazine.