Problem Statement: Geovisualization is a powerful tool for visualizing health data. The problem I addressed in my final project is how to take a large dataset containing influenza test results and utilize geovisualization to identify geographic trends within the data. This is of interest because from visualization, we can identify trends that may warrant more in-depth study.
Visualization: The first step in visualizing the data was to modify the data structure from the downloaded form to a matrix readable by Matlab with the date of test running down the rows and the location across the columns. The data in this form was used to create the animation of the test sites coming online. I then aggregated the data by month in order to identify geographic trends across the flu season. Through this visualization it is easy to identify hot spots for influenza in the US and potential sources of spread. For this visualization I utilized both graduated symbols and choropleth mapping. The graduated symbols represent the number of tests performed at a certain location while the choropleth color scheme represents the percentage of positive tests. Thus areas with high levels of influenza incidents will have a large, darkly colored circle. Finally, I created a cluster map which shows the location of the test sights that provided the data. This map is helpful as a reference when examining the month by month maps as it shows the locations without overlap.
Data: The dataset contains 458 locations over the period of a year ending in February 2016. The dataset contains over 63,000 records. The quality of the data is very good because it comes from a single source and I received the data directly from the source.
This animation shows the first test by each facility. It runs through the flu season, each location lighting up as soon as it performs the first test.
These six maps show the spread of influenza over the end of 2015 and the beginning of 2016. The symbols represent both number of tests from a particular location as well as the percentage of positive tests. A small, dark red circle represents a small number of tests, but a high percentage of them were positive. A large light circle represents a large number of tests but only a small percentage were positive for influenza. Of particular note are large, dark red circles. These maps identify a flu hot spot in the southwest, centering in Arizona.
This map aggregates locations that are clustered together. For example if there are several testing sites within a city, the city will be represented as a single circle with the number of sites located in the city displayed within the circle. If a testing site has no other sites within a close proximity, it will display as a small circle with no number.