Dr. Joe Champion
As GPS technology continues to advance, spatio-temporal data is collected by such devices at an increasingly higher sampling rate. This data can provide valuable insights on animal behavior patterns. However, processing hundreds of thousands of observations manually harbors potential for inefficiency. Furthermore, in-depth statistical analysis and visualization of the data would often require the use of additional tools. For the task of streamlining the processing, analysis, and visualization pipeline for cattle GPS collar data, we develop the animaltracker package in the statistical programming language R. With R Shiny, we construct a three-panel web application as the core feature of the package. The first panel allows for user-driven customization of data processing through filtering, elevation augmentation, and exporting. On the second panel, we visualize elevation time series and sampling rate by animal among others. On the final panel, we calculate statistical summaries for user-selected variables. More advanced visualization and analysis is possible through R functions included in the package, while data validation and outlier detection is also possible through a second Shiny application. By compiling these utilities in a framework designed for high-volume data such as the R environment, we provide a convenient platform that maximizes the efficiency of the pipeline.
Sukianto, Thea; Champion, Joe; Arispe, Sergio; and Mikesell, Dylan, "Animaltracker: Streamlining Spatio-Temporal Analysis and Visualization of High Sampling Rate Animal Data" (2020). 2020 Undergraduate Research Showcase. 184.