In one post titled My New Favorite Toy*** – Tableau, I mentioned my excitement for being able to start working on new visualizations using Tableau Software. Well, I’ve spent the last couple of weeks learning the tool and must say it is impressive. The biggest value that I’ve gained from using the tool is being able to take a large dataset and produce, within minutes rather than hours, meaningful visualizations of that data to the point that I can easily understand what’s truly going on with the data.
In fact, in the next few weeks, I should have a publicly available version of my first Tableau dashboard which I plan to publish to Tableau Public. For an advanced look at what I’ve got, see the dashboard below:
It may not be abundantly clear from the image above due to size constraints, but we have two visualizations at work here which can be updated by making filter selections in the center section of the dashboard. The dashboard is showing sold properties in the last 6 months in Mecklenburg County, NC (I live in Charlotte, NC).
On the left, we have a scatterplot. Along the x-axis, we have Sales Price while the y-axis displays heated square footage. These two variables are pretty strongly correlated in real estate, however there are so many more variables that go into real estate values so the scatter plot then uses color and symbols to further break out the data in the scatter plot. Notice, if you have 20-20 vision, that the color signifies the number of bedrooms and the symbol represents the type of deed that was conveyed when the property was sold.
On the right, we have a map of the Mecklenburg County area with all of the sales being mapped based upon the sold property’s geocoding.
This is a pretty high level view so the filters allow you to slice and dice the data. The first couple of filters, Area & Subarea, break Mecklenburg County up into the different realty areas used in our real estate market. Then, there are a couple of sliders where you can set the min/max settings for Heated Square Footage and Sale Price. Finally, there are three dropdowns where you can select a specific Property Type (Condo vs. Single Family Residential), Number of Bedrooms and Deed Type.
Above, you’ll see that we’re now drilled into area 5-4. Both the scatter plot and the map have refreshed to show only the properties for our filter selections. Based upon the geospatial map, you can start to see some clustering, mainly around number of bedrooms and deed type. Sales price isn’t visualized in the map at this time, although Tableau does make it an option.
For now, I’ll continue to play with this. I may change the symbols used for the deed type as the three most common are variations of an arrow which doesn’t make the visualization stand out so much.