Revealing Chemical\nTrends: Insights from Data-Driven\nVisualization and Patent Analysis in Exposomics Research

Understanding historical chemical usage is crucial for\nassessing\ncurrent and past impacts on human health and the environment and for\ninforming future regulatory decisions. However, past monitoring data\nare often limited in scope and number of chemicals, while suitable\nsample types are not always available for remeasurement. Data-driven\ncheminformatics methods for patent and literature data offer several\nopportunities to fill this gap. The <i>chemical stripes</i> were developed as an interactive, open source tool for visualizing\npatent and literature trends over time, inspired by the global warming\nand biodiversity stripes. This paper details the underlying code and\ndata sets behind the visualization, with a major focus on the patent\ndata sourced from PubChem, including patent origins, uses, and countries.\nOverall trends and specific examples are investigated in greater detail\nto explore both the promise and caveats that such data offer in assessing\nthe trends and patterns of chemical patents over time and across different\ngeographic regions. Despite a number of potential artifacts associated\nwith patent data extraction, the integration of cheminformatics, statistical\nanalysis, and data visualization tools can help generate valuable\ninsights that can both illuminate the chemical past and potentially\nserve toward an early warning system for the future.