Samples from flora, fauna, and soil and water have been gathered by the ERISS. The specific sample types can be clustered based on their ecological niche and purpose in the watershed, to contain the following groups: water, fish, molluscs, freshwater animals, freshwater plants, freshwater sediment, terrestrial animals, terrestrial plants, and terrestrial soil.

Have mining and milling operations at the Ranger Mine contaminated the surrounding Alligators River Region with radionuclides and metals.

– Where are the sample types of interest located in relation to the mine?

– How do the concentrations of radionuclides and metals change with distance from the mine?

The initial dataset was made available as a .xlsx file. The only manual process was to load the file into Excel and save as a .csv file; all other data cleaning was performed using an R script. Most of the data cleanup consisted of cosmetic and convenience changes, renaming variables and setting data types. We created a variable called sample_type to group samples by substrate, such as Terrestrial Plants or Fish. This is important for both our analysis and for the Shiny app. The dataset includes geospatial data using Easting and Northing in two different zones of a Mercator Projection. We used functions from the rgdal and sp packages to convert to lat-long coordinates, which we were then able to plot using ggplot2. A number of sample results were reported as below detection limits, in the form of "<value", where value is the detection limit rather than an actual measured value. We discussed several possibilities for managing these results. We found references that directly used the LOD as a result (highly conservative), that used 0 as the result (effectively loses results), or divided the LOD by a factor (2 or √2) for use as a result (could be considered arbitary). We elected to go the highly conservative route and use the LOD as the value.

The shiny app is an interactive map where the user can define inputs to map and view contaminant concentrations.

The first input is substrate, which the user can specify via a drop-down menu.

The second input is the date the sample type was collected. Since there are many different times of year represented for sample collection (ie the samples weren?t taken every year in October, for instance, but randomly across the year), the samples are selectable by year as the option. The user can specify the year via a slider input.

• Preston Phillips - Preston5789
• Scott Braley
• Chloe Stenkamp-Strahm