Data Visualization Tips: Think Beyond the Original Spec

Porter Thorndike's picture
 By | October 07, 2014
in Advanced Analytics, Business Intelligence, Dashboards, Data Discovery, data visualization, Data Visualization
October 07, 2014

Over the years, I‘ve spent a lot of time studying visualization and the best practices of displaying information.  I subscribe to the major agreed-upon principles (using the right visualization, minimalist chart formatting, flat color schemes, etc.) and find myself closely aligned with Stephen Few's principles. 

In this blog post, I'm going to share one of the most valuable principles that I live by when designing a dashboard or visualization: Think Beyond the Original Spec.

We are often presented with specs to produce a data visualization, such as “revenue over time” or “revenue by product”. 

Those sound like useful visualizations, but what’s missing from them?  Seeing how the revenue is trending doesn’t tell us much.  To add more value and perspective to your visualizations try:

  • Adding comparison (previous years) or target metrics
  • Analyzing whether there is a better metric (% change, variance)
  • Using percentages to expand comparison options

Here’s an example.  Using Information Builders’ WebFOCUS InfoDiscovery, I applied this principle to an analysis of trade data of the top GDP producing nations.  Simply displaying the total imports and exports by country for the previous 12 months is useful, but there are a couple of problems:

  • It’s hard to understand the trend since the trade numbers are large and may not fluctuate much
  • Comparison is difficult because the trade numbers differ radically between countries

Thinking beyond the original spec, what I really wanted to know was the percent change from month-to-month for these countries.  Using InfoDiscovery, I was quickly able to upload the trade data spreadsheet and calculate the change based on existing trade metrics. This produced some immediate benefits that can be applied anywhere you may be looking for a better metric for trending and comparison across products/customers/stores that have significantly differing total values.

  • The trend (percent increase/decrease) becomes much more pronounced
  • Comparison across all countries becomes possible; in this case, I used the Matrix Chart option in InfoDiscovery to visualize the data

I hope you found that useful. For more data visualization tips, and to see a live 30-minute demo of InfoDiscovery’s capabilities, join me for a Daytime Demo starting October 29.