Formatted Title
PFAS Data Overload: Updating Large Data Sets with Innovative Digital Tools
Background/Objectives
Sources of per- and polyfluoroalkyl substances (PFAS) contamination are difficult to discern on large sites which may have had multiple historic PFAS releases. Long-term sampling at a former industrial manufacturing site in Michigan led to the identification of PFAS as a co-contaminant of concern. Initial review of sampling results showed consistent PFAS levels, prompting targeted sampling efforts to evaluate site-wide patterns. Traditional data analysis tools such as data boxes and sample location maps were used to initially evaluate PFAS trends. A more robust profile evaluation was needed to refine the conceptual site model and provide direction for future sampling approaches.
Approach/Activities
Multiple visual tools were used to characterize PFAS to identify profile trends. This included tools to visualize PFAS fingerprints associated with sampling locations and transects that allowed comparison of fingerprints over a defined area to identify trends. Through this process, traditional methods to create these tools and evaluate the data were determined to be time consuming and inefficient when new data were collected. To make this process more efficient, a flexible and customizable web platform was developed to provide easily updatable visualizations and analyses of large data sets to facilitate efficient PFAS evaluations. Additional evaluations were built in using this larger data set including Mann-Kendall statistical test result diagrams, time series graphs, a data dashboard, and interactive maps including sample locations and results. Since the tools are linked to a central database, they can be rapidly updated when new data are reported from the laboratory. The project team used this innovative platform to evaluate trends across the site for PFAS and other constituents of concern as more data were collected.
Results/Lessons Learned
Initial evaluation of the data indicated the potential for the presence of more than one PFAS source. The practicality of the platform and streamlining of the visualization process created opportunities for cost savings throughout the project’s lifetime, making it a key resource for future work on this site. Rapid updates to the visualization tools as new data were collected provided the most significant benefit over traditional data review methods. The project team was able to make better informed decisions quicker because statistical analyses were readily available for review. Customizable location groupings and dynamically updating visualization tools facilitated project team review of PFAS patterns and trends. Challenges in creating the web platform and tools included optimization of data pre-processing and creating an efficient quality assurance/quality control workflow.