Data Science & Analytics
Data has a story to tell
Specializing in statistical and graphical presentation of biological and environmental datasets
Fundamental to efficient and effective use of data is the unbiased interpretation of those data, measured statements of statistical inference, and the intuitive presentation of analytical results. RTR’s data scientists are experts at analyzing complex datasets using sound statistical methodologies to generate insightful charts, graphs, maps, and data dashboards. Our team has expertise with multiple statistical and geospatial software packages, as well as a variety of data science programming languages.
- Power BI
- Power Query
- Visual Basic
Analytics that are innovative and defensible
Each dataset is unique requiring modelling approaches that are specific to the research questions being addressed. Our approach is not to force data into existing analytical frameworks, but instead, to build custom models best suited to each dataset. This requires RTR’s statisticians and data scientists to stay current on the latest statistical methodologies, and with this knowledge, work closely with each client to make sure that their needs are being met. We are not believers in the adage “status quo is the way to go”, instead we strive to reach new heights in innovation when it comes to the analytical approach used to answer research questions.
- Abundance, survival, and cause-specific mortality estimation using mark-recapture datasets
- Spatially balanced survey design, power analysis, and simulation studies
- Geospatial analysis of topographical and bathymetrical datasets, with a focus on riverine and estuary habitats
- Custom, dataset-specific models, including hierarchical Bayesian, nonparametric spatio-temporal, and autocorrelated data
- Construction of personalized custom R graphical and computational libraries
- Descriptive statistics for fisheries and wildlife datasets
Joint Mortality and Survival Model
Statisticians and Fisheries Scientists from RTR developed a state-space hierarchical Bayesian model to jointly estimate cause-specific mortality and survival of PIT-tagged juvenile salmonids across large spatial and temporal scales. The model incorporates recapture events of live fish and recoveries of tags from dead fish to generate more precise and accurate estimates of mortality and survival compared to traditional mark-recapture models. This methodical approach was recently published in the journal entitled Environmental and Ecological Statistics, with subsequent publications of results using this model in the Journal of Ecological Applications and Transactions of the American Fisheries Society.
Simulation Study on Parasitic Salmonid Diseases
Statisticians and Geospatial Analysts from RTR conducted a simulation study to determine if changes in river flows could be used to control the amount and quality of habitat used by aquatic annelid worms, a host species of myxozoan parasite found in salmonids in the Klamath River, USA. Simulations coupled a two-dimensional hydrodynamic model with a spatially explicit annelid habitat model to evaluate if desiccation events could be used to reduce annelid habitat and therefore reduce disease risks for salmonids. Results of this collaborative project between researchers from RTR, the U.S. Fish and Wildlife Service, and Oregon State University were recently published in the North American Journal of Fisheries Management.
Want to learn more?
Get in touch to learn how RTR can help meet your data, analytical, and research needs.