Python in Academia by Marco Bardoscia

400 views

Published on

Python in Academia by Marco Bardoscia

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
400
On SlideShare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Python in Academia by Marco Bardoscia

  1. 1. Python in Academia Marco Bardoscia International Centre forTheoretical Physics
  2. 2. Background > Theoretical (statistical) physicist > Economic systems (financial markets, economies) > Social systems (social networks) > Computational problems (SAT extensions) > Biological systems (metabolic networks)
  3. 3. Tools in Science “Using regular adhesive tape they managed to obtain a flake of carbon with a thickness of just one atom.” nobelprize.org
  4. 4. ComputationalTools > MATLAB > A single FORTRAN file (5000 lines of code) > C++ for data analysis > R
  5. 5. ComputationalTools > MATLAB > A single FORTRAN file (5000 lines of code) > C++ for data analysis > R … mostly true for small groups.
  6. 6. > General-purpose > Plenty of packages > Nicely extendable > Interactive But Python is:
  7. 7. > General-purpose > Plenty of packages > Nicely extendable > Interactive But Python is: … but many people in academia care about different things!
  8. 8. What people care about: > Is it fast (enough)? > Is it at least always faster than MATLAB? > What about my legacy code? (wrapping in Cython is hardly a choice) > Can I get nice publication-ready plots? (using defaults?)
  9. 9. What about reproducibility? If what I have done is interesting, someone else will do it from scratch. If they find the same results (modulo some tiny details) using an entirely different code, then we didn’t do anything terribly wrong…
  10. 10. > Up-to-date comparison with Matlab for the most relevant functions > Nice(r) defaults for Matplotlib plots > Simpler integration with compiled code What can be done?

×