The 4th Day of Analytics – Intersection

On the 2nd Day of Analytics Andrew Turner highlighted GeoIQ’s Merge capability, which allows users to consolidate the data and geographic features of multiple datasets into a single unified dataset.  Merge simplifies the analysis process by allowing several small datasets to be united and analyzed together,  but what if I have a very large dataset and want to map or perform analytics on a relatively small subset?  For example, let’s assume I am a regional marketing manager for Target and would like to analyze Black Friday Tweets mentioning Target that occurred within my region of responsibility, but I only have a dataset of Target tweets that occurred across the entire world.

This is where Intersection is useful.  GeoIQ’s new Intersection capability allows users to filter data by cropping out the features found within specified geographic areas or that overlap with the geographic features in another dataset.  Continuing with the above example and assuming that my region of responsibility is the state of Virginia,  the Intersection capability would allow me to extract any tweets from the worldwide Black Friday Tweets dataset that are contained with Virginia boundaries (counties, zip codes, census tracts, etc.).  This operation results in a new GeoIQ dataset comprised of only the tweets that I care about, which I can then map and analyze further.

Let’s walk through the workflow. I will start with my worldwide Black Friday Tweets dataset:

In GeoIQ I will choose to perform an Intersection against the Black Friday Tweets dataset, at which point I will be prompted to select a dataset to intersect with.  I will choose the Virginia County boundaries dataset shown below:

After kicking off the Intersection I will be asked to provide metadata for the resulting dataset, after which I can map the new dataset and perform additional analytics against it.

The above map shows my Virginia-based tweets as proportional symbols sized by the number of followers for the user who tweeted, all of which is overlayed atop the Virginia boundaries that were used to intersect the original worldwide tweets dataset.  The new Virginia dataset could be merged with additional datasets if my region of responsibility spans outside of Virginia, or it could be aggregated up to Virginia boundaries if I wish to analyze at a regional rather than point level.

GeoIQ’s Intersection capability is a powerful feature that allows anyone to perform a spatial filter with only a couple clicks through an intuitive workflow, and it truly shines when combined with additional GeoIQ analytics.  In a couple days one of my colleagues will be demonstrating how Intersection and another GeoIQ analytic capability can be chained together to further simplify data discovery.  Please check back!