SciPy Get Started
This page walks through installing SciPy, checking that the install worked, and understanding how its subpackages are imported.
Installing SciPy
- pip install scipy - the standard way from PyPI
- conda install scipy - if you manage packages with Anaconda or Miniconda
- python -m pip install --upgrade scipy - to upgrade an existing install
SciPy depends on NumPy, so installing SciPy will pull in a compatible NumPy version automatically if one is not already present. You do not need to install NumPy separately first.
Checking Your Installation
Check the version
import scipy
print(scipy.__version__)Importing scipy on its own only gives you a small set of top-level items, such as the version string and testing helpers. It does not automatically load scipy.optimize, scipy.sparse, or any other subpackage.
Note: Calling scipy.optimize.minimize(...) right after import scipy will raise an AttributeError in many SciPy versions. Always import the subpackage you need directly, for example from scipy import optimize.
Import and use a subpackage
from scipy import optimize, constants
print(constants.speed_of_light) # 299792458.0
result = optimize.minimize(lambda x: (x - 3) ** 2, x0=0)
print(result.x) # array([3.])A Typical SciPy Workflow
- Import NumPy for arrays and basic math
- Import the specific SciPy subpackage(s) your task needs
- Prepare your input data as NumPy arrays
- Call the relevant SciPy function
- Read the result from the returned object's attributes, such as .x or .fun
Note: Inside a Python shell, help(optimize.minimize) or dir(optimize) is a fast way to discover what a subpackage offers before you go looking through the documentation.
Exercise: SciPy Get Started
Which command installs SciPy using pip?
