A Better Graph - From Start to Finish

At Arcadia, we spend a lot of time thinking about how to make data useful to our customers – whether in Arcadia Analytics, in our Data Gallery, or even in a simple graph.  It’s easy to make a graph – but deceptively difficult to make that graph useful to its intended audience.   Here is our simple guide to making a better graph in 11 steps.

Step 1: Pick the right starting point.

Be sure to choose the right type of graph for the data you have and the story you want to present.   In our example, we’ll use a bubble chart. This kind of chart is a great way to convey complex data with X, Y, Z, and Color dimensions. You can also add vectors for a 5th piece of data!

Step 1 - Choose the right starting point

Step 2: Scale your data.

Apply scales (logarithmic, power) to let your data breathe and better understand trends.  Be sure to tell your readers, though.

step-2-scale-your-data

Step 3: Color your data.

Use color to help your reader understand different parts of your graph.  Multiple graphs about the same topic?  Be consistent with your colors to avoid confusion.  And when using transparency and overlays, choose colors that mix well (i.e. not complementary colors).

Step 3 - Color Your Data

Step 4: Style opacity and borders.

Increase transparency for overlapping data. Use borders to make sharper edges, which make it easier to find distinct shapes in data clusters.

Step 4 - Style Opacity and Borders

Step 5: Zero out your axes. 

Unless you have a compelling reason, force your axes to zero. If you don’t – make it obvious where your axis is starting. In this case, the negative portions of the axes were superfluous.

Step 5 - Zero Out Your Axes

Step 6: Remove discretionary gridlines.

How important is it for your user to pinpoint the X-Y coordinates of a data point? If it’s not that important, remove some of the grid noise.

Step 6 - Remove Discretionary Gridlines

Step 7: Fall in love with fonts.

Generic fonts have come a long way (Calibri > Arial).  BUT – you can differentiate your visual with one of thousands of beautiful, unique fonts.  Google has a good resource for fonts.

Step 7 - Fall in Love with Fonts

Step 8: Label Your Axes

Add labels to your axes, including any scales applied and the units being measured.  Put a simple explanation at the top of the graph to remove ambiguity.

Step 8 - Label Your Axes

Step 9: Create a great legend.

Labels on the axes are the most intuitive, and you should generally keep them there, but sometimes you may want to show those labels in a legend that conveys the information in a unique – or more informative – way.  In this example, the use of the legend supports a more comprehensive explanation of how to interpret the bubbles.

Step 9 - Create a Great Legend

Step 10: Label your series.

While simple series labels work well enough, you can use this opportunity to embed some of your findings and analysis directly into the graph.

Step 10 - Label Your Series

Step 11: Call out your data. 

Add some callouts to help tell your story and highlight interesting pieces of information.  In our example, the called-out bubbles are styled opaquely, to stand out from the other (transparent) bubbles.

Step 11 - Call Out Your Data

Each step by itself is simple, but in combination they create a much more compelling graph.  To check out some of the ways we use these ideas in our data visualization work, visit the Arcadia Data Gallery.

Nick Stepro

Nick Stepro

Nick Stepro is the senior vice president of product management at Arcadia, where he leads the design of the next wave of advanced healthcare analytics applications – including Arcadia Analytics, which has been praised as having one of the best user interfaces in the industry.  He has worked with large health systems and payers to design and execute on innovative clinical integration and business intelligence strategies to drive improved health outcomes and reduced system costs.

Nick believes in good design and data visualization. When combined with focused expertise in analytics, healthcare and business process, the results are intuitive data-driven applications that empower users to dramatically improve the way they run their businesses.   His data visualization work has been covered on NPR, U.S. News and World Report, Medical Ethics Advisor, and elsewhere.  Becker’s Health IT and CIO Review recently named him one of “31 Health IT and Revenue Cycle Whiz Kids” to watch.  He has spoken at Medcity CONVERGE, AMIA, and HIMSS and has been a guest lecturer on data visualization at Georgia Tech.   In December 2016, he was the closing speaker at the CCO Oregon Cost of Care conference.

Specialties: Data Visualization, Product Management, Healthcare Data, Population Health Management, Solution Architecture, Account and Relationship Management, Business Development, Marketing, Graphic Design, HIE, EHR, HIE, PCMH, ACO, P4P and many more important acronyms.

December 15, 2016
,

The Final Year - A Data Visualization of End of Life Care

Nick Stepro’s data visualization The Final Year was highlighted on NPR, U.S. News and World Report, and elsewhere.  It tells the story of the final year of life for 2,398 patients who died within a five-year span.   What services did each patient receive?  What does it mean to die in a hospital versus at home?   How do our end of life decisions impact our quality of life and our cost of care?

Explore the data visualization