This is part of my continuous experiment to better visualize spatial data. An earlier of post is here:
When I was working on visualizing the disease cases and ratios in California, I wanted to see the spatial distribution of the disease cases. And I used Pareto method to help create contiguous quantiles (equal portions in population) on the map, from which resulted the above picture. More examples are as follows. We can use the Pareto measure to create various partitions as we wish.
This is a quasi-quartile partition of the populations based on county data. We see that Los Angeles county got about one quarter of the total population. No wonder it is the most populous county in the state.
This picture shows that the north of San Francisco Bay Area is quite sparse with only 5% of the population. About a quarter of the population live between San Francisco and Los Angeles. So, we got good insights into the population distribution in California.
The Visual Pareto Approach
In the above examples we use the North-South approach. The county population data in California is scraped from here. The county latitude data is generated from Tableau but has to be exported to an Excel file and re-imported into Tableau, because we couldn't directly reference the latitude(generated) in our calculation. (What a waste! Thus I created a request for the feature. Seems other people already asked for it long time ago. Please vote for them if you believe it's useful.)
Assume Pareto cumulation will be computed North-South along the latitude axis. We can create a regular Pareto chart as below.
The Pareto chart may tell us a few things. But it's not so intuitive about the population distribution. In combination with map partition, we create something that's easy to visually interpret.
The main steps to create the above partitioned maps are:
1.Create a table to get the latitude of all the data points (counties here). Export it to Excel and import it again into Tableau. Then the latitude data is made referenceable.
2.Use this data source as the primary to blend with the original data set from which we can get measures (such as populations or disease cases).
3.Put County in the Detail shelf and sort it by AVG(Latitude) descendingly from North to South.
4.Drag Population to the Detail shelf and set Sum(Population) to quick table calculation of Running Total. Set the table calculation to compute using County. In this step we may create a few more measures such as percent of total to show in tooltips.
5.Create partitions to color the map. (The calculation is based on the Pareto method.) Then, drag the Quartiles (in this example) to the Color shelf.
Voila. The workbook can be download here.
Add a comment