Demographic Correlation

Overview

The Demographic Correlation insight panel is an analytics that allows you to view the pearson correlation coefficients, a standard statistical metric of correlating two data series, for one or more demographics and sales of one or more items.  The power of this insight panel lies in the ability to view multiple correlations quickly to identify which ones may be positive or negative drivers of incremental sales. 

Anatomy of the Demographic Correlation Insight Panel



Insight Gear Customization Options

Gear options for this panel are numerous as it provides you all the options to customize your analysis appropriately.  As screenshot is shown below, see after the jump for a listing and explanation of each gear option selector.

The first three options are fairly common to insight panels :

Retailer - you need to choose a retailer to pull sales from.  You can only choose one at a time here.  To do analysis of multiple retailers, just create another insight panel and side-by-side.

Sales Metric - you need to determine what sales metric you're going to correlate to one or more demographics.  The most commons choice here is retail sales.

Time Period - you need to choose a time period to pull sales for.  We recommend a fairly long time period as correlations become more statistically significant the larger the data size there is to draw upon.  We specifically suggest you avoid time periods less than 12 weeks as the sample size is almost always too small and erroneous resultst and conclusions become likely.

Demographic(s) - Here is where you can select he demographic variables you wish to correlate.  While the gear provides for many options, we recommend the % of population variables as they make the most sense.

Miles Around Store - this is used to tighten the distance used around each store to get a broader or more concentrated view of the local demographics.  5 miles is a  good value.

Filter on Custom Segment - this option allows you to utilize custom store segments you've created previously to limit analysis to within the segment.  This can be very handy if you want to view correlations of demographics, but only within a subset of the population limited based on another demographic. 

Filter Correlations - this options allows you to only return correlations in certain pre-defined segments of strength, allowing you to ignore weak correlations, for example.

Single Item / Category Type / Category / Custom Item Group - these options which we've grouped together are standard options for limiting the items being evaluated.  Its useful to note that the more items included in your analysis, the lower the strength of the correlations - this is because there may be offsetting strengths in the underlying items which will dilute the result.

Custom Region / State - these options allow you further limit the geography of the analysis. Unlike the custom segment, this is only based on geography.

Control Other Insights - this option can activate this panels ability to make other panels in the same view react based on your interaction with this panel.   

Table of Contents