Revealing Chemical Trends: Insights from Data-Driven Visualisation and Patent Analysis in Exposomics Research#

Authors#

Dagny Aurich, Emma Schymanski, Flavio De Jesus Matias, Paul A. Thiessen, Jun Pang

Abstract#

Understanding historical chemical usage is crucial for assessing current and past impacts on human health and the environment and informing future regulatory decisions. However, past monitoring data is often limited in scope and number of chemicals, while suitable sample types are not always available for remeasurement. Data-driven cheminformatics methods on patent and literature data offer several opportunities to fill this gap. The chemical stripes were developed as an interactive, open source tool for visualising patent and literature trends over time, inspired by the global warming and biodiversity stripes. This paper details the underlying code and datasets behind the visualisation, with a major focus on the patent data sourced from PubChem, including patent origins, uses, and countries. Overall trends and specific examples are investigated in greater detail to explore both the promise and caveats that such data offers in assessing the trends and patterns of chemical patents over time and across different geographic regions. Despite a number of potential artefacts associated with patent data extraction, the integration of cheminformatics, statistical analysis, and data visualisation tools can help generate valuable insights that can both illuminate the chemical past and potentially serve towards an early warning system for the future.

Keywords: Chemical Stripes, PubChem, Cheminformatics, Data Visualisation, Exposomics, Early Warning System, Patent Analysis