Project based Python teaching: Lessons from the two body problem and beyond

Teaching programming to science master students with different programming experience can be difficult. Allowing more advanced students to work on a project can help providing challenges for each student. The talk describe

Transcript

1. Project based Python teaching: Lessons from the two body problem and beyond Python in Astronomy - Leiden - 10.04.2017 Lucia Klarmann – University of Amsterdam These slides use elements from SlidesCarnival
2. API 2017 CWI 1989: Python
3. Let's teach python to all master students! But what about the students who already know python?
4. But what about the students who already know python? initial test project course
5. Project Interesting and challenging, but not too difficult Allow students to choose their focus Astronomical context Highlight python as a programming language Prepare students for their master project Within the TA's area of expertise... c
6. Two body problem
7. Interesting and challenging, but not too difficult Allow students to choose their focus Astronomical context Highlight python as a programming language Within the TA's area of expertise... cc c
8. Interesting and challenging, but not too difficult Allow students to choose their focus Astronomical context Highlight python as a programming language Within the TA's area of expertise... c
9. Interesting and challenging, but not too difficult Allow students to choose their focus Astronomical context Highlight python as a programming language Within the TA's area of expertise... c
10. Task: Solve the two-body problem problem program integrators setups use/expand code topic of your choice
11. Task: Solve the two-body problem modular code readability counts performance and unit testing documentation
12. Task: Solve the two-body problem write it up in a final report make plots and animations evaluate methods and analyse results
13. Task: Solve the two-body problem problem program integrators setups use/expand code topic of your choice write it up in a final report performance and unit testing documentation make plots and animations evaluate methods and analyse results modular code readability counts
14. ... and beyond
15. F. Backs Results in a nutshell
16. Euler, leapfrog and Runge-Kutta comparison energy Angular mom. time D. van Eijnaten
17. Introducing variable, individual timesteps F. Backs
18. Introducing variable, individual timesteps F. Backs
19. It is hard to get better than SciPy F. Backs
20. Simulating the solar system over 5 Gyr? D. van Eijnaten
21. The project worked very well The students worked well and were motivated The students used online python resources The code was modular and well readable The finally reports were all good or very good
22. There is always room for improvement The quality of the documentation varied Unittest coverage could have been higher Not everything ended up in the final report The reason for these problems was lack of time The connection to current research was minimal There were some difficult debugging sessions
23. Yes, the students liked the project The exam was very python specific The two body problem is not very glamorous Time management was difficult The overall feedback was very positive They enjoyed the freedom during the project
24. Yes, the students liked the project The exam was very python specific The two body problem is not very glamorous Time management was difficult The overall feedback was very positive They enjoyed the freedom during the project
25. Yes, the students liked the project The exam was very python specific The two body problem is not very glamorous Time management was difficult The overall feedback was very positive They enjoyed the freedom during the project
26. c write test
27. c write test
28. Teach python to all (astronomy master students) Keep talking Try the 2 body problem Independent work write test