This question is probably at least applicable for those of you that work in a commercial setting.

I am working in a rather general way on process improvements with regards to communicating (or not communicating) with software engineers about statistical work product. In particular I would love to hear what anecdotal experiences, unsolved problems, best practices, and solutions you have encountered in the process of communicating data science, machine learning, or statistical results and processes to someone or to a team.

To state this as a question: What best practices do you use, if any, and is this process as effective as you would like and why or why not?

I am also interested to know if you require a level of statistical education from your software developers and are you required to develop any computer science related education in order to participate in your process.

  • $\begingroup$ Maybe you should ask do you require a level of floating point numerical mathematical software, i.e., numerical analysis, education from your software developers? It's quite apparent that many data scientists have no such education worth a damn. $\endgroup$ Sep 10, 2015 at 20:13
  • $\begingroup$ Thanks! I added the possibility of requiring some computer science related education. I intend to provide a complementary question in StackOverflow shortly $\endgroup$ Sep 10, 2015 at 20:24


Browse other questions tagged or ask your own question.