Data Science Projects

Repositories for these projects are available on GitHub at andrewguenthner/datascience

Featured Projects:

Baltimore’s been in the news lately! See how the metro area has fared in terms of the health habits and health risks of its people.

This interactive Web tool automatically analyzes and classifies time-series trends in key health indicators for over 250 US metro areas. Visualizations enable you to see trends for each indicator, compare metro areas on the basis of the absolute value of every indicator in every year, and let you compare metro areas according to trend class (which ones are improving rapidly vs. which ones are getting worse but stabilizing, for instance).

Want to see how players match up for the next Bay Area disc golf league games? We built just the tool.

Machine learning and AI don’t just have to be world-changing technologies, they can be fun to see in action. I teamed up with Luiz Celeste, an avid disc golfer and ML practitioner, as well as Don Cushing, a great web designer, writer, and data scientist. We decided to make the games a bit more interesting by letting players see how our machine learning model, which takes into account both weather and time-series factors, rates them against their competitors. Validation tests show we can predict player scores to within 5 strokes 80% or more of the time.

Can you spot Hurricane Florence? (look for the green streak connecting West Africa and the southeastern US)

The world’s weather in 2018: see how products from supercomputer analyses, which are newly available to the public, make great data visualizations. The rainfall from Hurricane Florence is visible as a green streak connecting West Africa and the southeastern US in the image above.

Other Projects:

  • The world of health care: unique data visualizations showing correlation and time-series of key health care indicators for countries around the world
  • Hansen Solubility Parameters: prediction tool that’s highly useful for biologists and chemists. This project has been published in a scientific journal.
  • Predicting Properties of Mixtures and Biomaterials: More projects that are especially useful for materials scientists.
  • The team I lead at OpenOakland (a brigade in the Code for America network) has developed and deployed tools for Oakland Art Murmur, such as this interactive map.