Data Science

We are passionate about data science and want to help your research or business! We can help with the following:

  • Selection of research design and data collection methods. Whether questionnaires, surveys, observational, or data from documents and records, we can recommend and help with a research design and data collection method that affords you the best opportunity to address your research or business objectives.

  • Research data organization, secure storage, and sharing. We can help administer the hosting of your data in the cloud, along with all R code, workflows, and Shiny apps you need. We can help with Amazon Web Services (AWS), Google Cloud Computing (GCP), or RStudio Connect.

RStudio Connect is the cloud based service to host all types of R products. We have an example of an application we are currently hosting on RStudio Connect here. Jen Underwood has a nice summary article on RStudio Connect

AWS provides a summary of running R on AWS here.

Empirical Path provides a good example of running RStuido/RShiny on the Google Cloud Platform

  • Research data processing, manipulation, and visualization. We use R and/or Python to wrangle data. Often, data wrangling, cleaning, and preparation for analysis is the most time consuming part of data analysis. This process also greatly influences the type of tools used for visualization.

  • Selection of data mining, machine learning, and statistical analysis. We use R and/or Python to conduct nearly all data mining, machine learning, and statistical analysis. We rely heavily on the caret, _sjPlot, and R stats package.

  • Study execution, results interpretation, and visualization. Complete methods description, data analysis, explanation of results, with tables figures and legends will all be prepared in an Rmarkdown document, which can be exported to HTML, PowerPoint, Word, or PDF. All code is made available, in support of reproducible research methodology. We will work with you in a step wise manner to complete the data science project. In general, we follow the REASON method outlined by Damian Mingle here.