An application for automation and streamlining of workflow for modeling of time-series data
Expand usability of existing water-quality software more broadly to other time-series datasets
The USGS regularly estimates water-quality conditions in U.S. waters by developing statistical relations among continuous in-stream sensor readings and results from laboratory analysis of discrete water-quality samples. Despite widespread use of this approach, the procedures used to develop, quality assure, and publish these relations vary by individual authors. The project proposes to publish an R package that improves the consistency and speed with which these relations are established, improve capacity to access published model data, and implement more sophisticated modeling techniques specific to datasets including censored values (such as <0.1 mg/L). This package would also be well-suited to a wider range of non-water environmental parameters characterizable in time-series by ordinary least squares or survival regressions based on uncensored predictor variables. This project will utilize existing tools and websites to provide an immediate benefit to many stakeholders across multiple disciplines that employ these approaches and improve access to these data.
wren: water-quality regression tools wren: water-quality regression tools
Expand usability of existing water-quality software more broadly to other time-series datasets
The USGS regularly estimates water-quality conditions in U.S. waters by developing statistical relations among continuous in-stream sensor readings and results from laboratory analysis of discrete water-quality samples. Despite widespread use of this approach, the procedures used to develop, quality assure, and publish these relations vary by individual authors. The project proposes to publish an R package that improves the consistency and speed with which these relations are established, improve capacity to access published model data, and implement more sophisticated modeling techniques specific to datasets including censored values (such as <0.1 mg/L). This package would also be well-suited to a wider range of non-water environmental parameters characterizable in time-series by ordinary least squares or survival regressions based on uncensored predictor variables. This project will utilize existing tools and websites to provide an immediate benefit to many stakeholders across multiple disciplines that employ these approaches and improve access to these data.