TEACHING DATA SCIENCE TO SCIENTISTS David Nicholson, Ph.D. Emory University Sober lab 

THE NEED • All fields of science now flooded in digital data • Scientists need data science training • E.g., graduate students • To better deal with their data • To have hard skills useful in both in academia and out • Not all fields of study / programs have recognized this need • Astrophysics v. neuroscience • The preqrequisite pipeline • https://github.com/SarTaku/Emory-Data-Science-Classes 

A SOLUTION • Atlanta BEST program: Broadening Experiences in Scientific Education • http://www.best.emory.edu/ • Program manager Tami Hutto, [email protected] • Accomplishments over the last year: • Atlanta BEST Data Science club • Atlanta BEST Data Science Github organization • Short course on scientific computing and data scientists 

ATLANTA BEST DATA SCIENCE CLUB • http://www.best.emory.edu/resources/data-science-club.html • Monthly meetings • Organize through Slack (datascienceatl.slack.com) and listserv • 5-10m presentations on topics of interest • git + GitHub -- https://github.com/Atlanta-BEST-Data-Science/github_help • SQL • Kaggle • Panel discussions with alumni and local experts in data science • @PyDataATL https://www.meetup.com/PyData-Atlanta/ • AniData http://anidata.org/ https://github.com/anidata 

ATLANTA BEST DATA SCIENCE GITHUB ORGANIZATION • http://www.best.emory.edu/resources/data-science-club.html • Repos based on talks • Git + GitHub help • List of contacts: https://github.com/Atlanta-BEST-Data-Science/BEST-contacts • that taught all club members basic git + Github skills • shows professors + administrators at Emory, GA Tech, GSU the hunger for and need to provide these skills 

ATLANTA BEST DATA SCIENCE SHORT COURSE • https://github.com/Atlanta-BEST-Data-Science/BEST-data-science-short-course • Based on: • Software Carpentry for git, command line, Python • https://software-carpentry.org/lessons/ • Jake Vanderplas’ Data Science Handbook • https://github.com/jakevdp/PythonDataScienceHandbook • Elegant SciPy, great book by @jnuneziglesias @hdashnow and @stefanvdwalt • https://github.com/elegant-scipy/elegant-scipy • Class project • Create a new Github repo with jupyter notebook, using newfound data science skills: git + command line + Python + TDD + numpy scipy matplotlib 

TAKE HOMES • Scientists need data science education! • Yes, there are open sourced tools that make it easy (easier) to address this need • Textbooks • Github itself! Student Developer pack • PyData, related orgs and meetups • You can help address this need. Learn from our trial and error! • http://www.best.emory.edu/resources/data-science-club.html • http://www.best.emory.edu/resources/data-science-club.html • https://github.com/Atlanta-BEST-Data-Science/BEST-data-science-short-course I’ll tweet these slides after the talk so you can get the links  @nicholdav