King County Pollution Distribution

I made this series of maps for a project in ENVS-420 (GIS Analysis and Modeling). The premise of the project was to take US Census Bureau data and the locations of known pollution sites to analyze how BIPOC and impoverished communities are affected by pollution.

To derive the data shown on these maps, I first downloaded census data from the portal online, and formatted it to be importable to ArcGIS Pro. Then, I created a model that uses census polygons, locations of pollution sites, and parks/waterbodies as inputs. The model splits some existing census blocks in half, so each block is updated with a new area and population fields (assuming uniform population density). Waterbodies and parks are excluded on the area field, since a negligible amount of people live there. The model outputs a census blocks dataset with many fields, such as whether a block is in/out the county, the percentage of people in the block living in poverty, etc. More details on the analysis visible below (Fig. 4).

The data reveals the gruesome reality that BIPOC and impoverished communities are disproportionately exposed to pollution in King County. These inequities stem from issues hundreds of years old, yet are still affecting people today.

Fig. 1: Shows the distribution of BIPOC and impoverished communities in relation to pollution sources in King County. Darker colors mean higher concentrations of BIPOC and/or impoverished people. Pink rings represent 4 kilometer radius “pollution zones” around known pollution sources. This data suggests that BIPOC and impoverished people are disproportionately exposed to pollution in King County.

Fig. 2: Comparison of classification methods using U.S. Census Bureau income data. Classification refers to how you break up your data classes. This will affect what data falls into which class and ultimately which color your data is assigned to. This figure shows how big the impact that this can have on the reader’s perception of the data. Every map is symbolized by the total number of impoverished people per polygon (impoverished is defined as having income less than two times over the poverty line). However, each map paints a drastically different picture of poverty in King County.

Fig. 3: Shows the difference that normalizing data can make. Normalizing means adjusting data to account for differences in some property. In this example, Map 1 shows polygons symbolized for total BIPOC population, and Map 3 shows polygons symbolized for BIPOC population density. The problem with displaying the total population per polygon with different sized polygons is that there tend to be more people living in larger areas. Normalizing this data for area (or calculating population density) simply means dividing the total population by the area where that population was gathered. As seen in Figure 2, this can have a drastic impact on the reader’s perception of a map!

Fig. 4: Model used to conduct analysis. The model takes three input datasets of pollution sources, census block data, and people-free zones (in our case, water and parks). It splits the census blocks into zones inside/outside the buffers, performs calculations of the new area of those polygons, and adjusts the population values assuming a uniform population density. This updated dataset is the final output of the model.

Page updated

Google Sites

Report abuse