Last week Andrew C. Oliver, manager of technical enablement at Lucidworks and InfoWorld columnist, shared his 12 New Year’s Resolutions for Your Data. He has some good ideas about better handling your data (his introduction via the mainframes-through-Hadoop-journey made me smile). The main focus of the article talks about ways to better understand how data flows through your organization, and how to improve the use of the data via things like accessibility. It was his last suggestion, to “make it visual,” that inspired us to create our list of ways that you can make your big data visual.
The list below isn’t all encompassing but does give you a menu of choices for the necessary step of visualizing your big data to gain actionable insights for your organization.
Line charts connect discrete values to one another in a manner that effectively illustrates the shape of change over time. Use a line chart when you want to visualize measures within a time series.
If plotting multiple measures, consider adding points to the lines to facilitate easier comparison at a given point in time.
Bar charts are visually weightier than points and lines. Bars convey quantitative meaning not only by end points, but also by length. They allow readers to compare individual values to one another. You’ll want to use a bar chart for measures that are nicely grouped into different categories.
Bar Chart Variations
Use grouped bars when you want to compare different measures by the same category.
Use stacked bars to illustrate proportions of (up to) a few categories as they relate to a whole. You’ll want to put the most important category along the axis to help readers quickly see the information you want to emphasize.
Those are the basic options. We agree, Andrew is right, “People like charts—lots of charts and pretty lines.” But why be limited to only charts and lines? Below are a few examples of other types of visualizations.
Scatter plots are arguably one of the simplest, yet most effective visualizations. They enable readers to plot values as they relate to the scales of the X and Y axis. You would use scatter plots when comparing two quantitative scales or for comparing a category to a quantity.
A calendar heatmap augments a time series analysis using color intensity. You can use this chart type to quickly illustrate the differences between smaller and larger quantities across a calendar year.
A map offers a familiar, intuitive canvas for presenting measures with geographic attributes. For a fixed analysis using standardized measures (averages, ratios, etc.) use a choropleth.
An interactive map gives the reader multiple layers of data to explore and can be valuable for uncovering insights at lower levels of detail.
A Sankey diagram is a multifaceted type of flowchart in which the width of a path is proportional to the quantity. This diagram also demonstrates the different patterns in an event sequence. Given an identified starting point and goal, the chart shows the events that occur between the two. For instance, use this chart type to better understand clickstream data.
A chord diagram displays the relationships among categories. The connection between two values shows what they share in common. You should consider using this chart type with two attributes that have low cardinality.
While selecting visuals is only one step in the big data journey, hopefully this article helps round out Mr. Oliver’s list of steps in the process of collecting, managing and visualizing your big data.The list above is only a subset of the many visual types Arcadia Data offers out of the box. Download Arcadia Instant today to explore these and more, and find out how you can uncover valuable insights from your big data.