Got the SciPy packages installed? Wondering what to do next?
“Scientific Python” doesn’t exist without “Python”. SciPy skills need to build on a foundation of standard programming skills. While Python itself has an official tutorial , countless resources exist online, in hard copy, in person, or whatever format you prefer.
Just remember to have fun, make mistakes, and persevere.
Where to write¶
Jupyter notebooks combine code, markdown, and more in an interactive setting. They are an excellent tool for learning, collaborating, experimenting, or documenting. Notebooks can run on your local machine, and MyBinder also serves Jupyter notebooks to the browser without the need for anything on the local computer. For example, MyBinder Elegant Scipy provides an interactive tutorial.
Jupyter runs by calling to IPython behind the scenes, but IPython itself also acts as a standalone tool. A command-line of individual statements and returned values, IPython is useful for debugging and experimenting.
Code Editors and IDEs (Integrated Development Environments) facilitate the writing of scripts, packages, and libraries. These tools handle projects, like SciPy itself, that start to grow larger and more complicated. Separate files can hold frequently used functions, types, variables, and analysis scripts for simpler, more maintainable, and more reusable code.
Code editors run from minimal, like Window’s Notepad, to the fully-featured and customizable, like Atom , Visual Studio Code , or PyCharm. Features include syntax highlighting, the ability to execute code, debugging tools, autocompletion, and project management.
Need to test if the packages got installed? Type these lines at an IPython
prompt, or save in a
*.py file to execute:
import numpy as np print("I like ", np.pi)
For testing the SciPy library and Matplotlib, here’s a fun Easter egg:
from scipy import misc import matplotlib.pyplot as plt face = misc.face() plt.imshow(face) plt.show()
Each package has official tutorials:
But the best way to learn is to start coding.
Stuck? Need help?¶
Getting errors that you can’t figure out?
Start by looking at the error message. Yes, error messages are often intimidating and filled with technical detail. However, they can often help pinpoint the exact location in code where things go wrong. This is often most of the battle.
Unsure of how to use a particular function? In Jupyter and the IPython shell, call up documentation with:
import numpy as np np.linspace?
or for viewing the source:
import numpy as np np.linspace??
? works on both functions and variables:
a = "SciPy is awesome ;)" a?
Try searching the Internet and sites like StackOverflow to see if others have encountered similar problems or can help with yours.
If you think you have truly encountered a problem with SciPy itself, read the page on Reporting Bugs.