Data Science Skills

Import and Install Python Packages for Science

Learn about managing Python packages in your code.

What is a Python Package

In Python, a package is a bundle of pre-built functionality that adds to the functionality available in base Python. Base Python can do many things such as perform math and other operations. However, Python packages can significantly extend this functionality.

You can think of a Python package as a toolbox filled with tools. The tools in the toolbox can be used to do things that you would have to otherwise hand code in base Python. These tasks are things that many people might want to do in Python, thus warranting the creation of a package. After all, it doesn’t make sense for everyone to hand code everything!

For example, the matplotlib package allows you to create plots of data. Since most of us create plots routinely, having a Python package to create plots makes programming more efficient for everyone who needs to create plots.

Open Source Python Packages for spatial and time series data

There are many different packages available for Python. Some of these are optimized for scientific tasks such as:

• Statistics
• Machine learning
• Using geospatial data
• Plotting & visualizing data
• Accessing data programmatically

and more! The list below contains a few core packages that are often used by scientists as examples.

• os: handle files and directories.
• glob: create lists of files and directories for batch processing.
• matplotlib: plot data.
• numpy: work with data in array formats (often related to imagery and raster format data).
• pandas: work with tabular data in a DataFrame format.
• rioxarray: work with raster (image and arrays) data using XArray and rioxarray approaches.
• geopandas: work with vector format (shapefiles, geojson - points, lines and polygons) using a geodataframe format.

Where Do Packages Live On Your Computer?

Packages are organized directories of code that can be installed and then imported to your code file (e.g. .py script, Jupyter Notebook file).

When you install a package, you may be wondering, where does it go? If you are using mambaforge on a Mac, the packages that you install are located in your mambaforge directory under envs (e.g. /home/username/mambaforge/envs/).

When you install a package into the conda environment of your choice. For example, geo-python that you installed in this tutorial series will end up in the /home/username/mambaforge/envs/geo-python folder.

Once packages are installed in your Python environment (e.g. geo-python conda environment), you can call them in Python at the command line, in a script (.py file), or in a Jupyter Notebook file.

You have to explicitly call and load (i.e. import) each package that you want to use in your script (.py file) or Jupyter Notebook file, in order for the functions (or tools) in that package to be available for use in your code.

Python Packages Can Contain Modules

Packages can contain many modules (i.e. units of code) that each provide different functions and can build on each other. For example, the matplotlib package provides functionality to plot data using modules, one of which is the commonly used module called pyplot.

Every Python package should have a unique name. This allows you to import the package using the name with the import command.

For example, the command below imports the matplotlib package.

import matplotlib

Now you can create a plot but notice that you have to call matplotlib as a full word each time you run a plot command below.

import matplotlib

# Sample data
x = [1, 2, 3, 4, 5]
y = [2, 4, 6, 8, 10]

# Create a basic plot
matplotlib.pyplot.plot(x, y)

# Add labels and title
matplotlib.pyplot.xlabel("X-axis")
matplotlib.pyplot.ylabel("Y-axis")
matplotlib.pyplot.title("Simple Matplotlib Plot")

# Show the plot
matplotlib.pyplot.show()

What is a Module in a Python Package?

Packages often have modules. A module is a set of related functionality that lives within the package.

For example, pyplot is a module within the matplotlib package that makes it easier to quickly set up plots.

You can import a specific module like pyplot by first calling the package name and then the module name - using . to separate the names like this:

import matplotlib.pyplot

You can also import the module using an alias or short name, such as plt for matplotlib.pyplot.

import matplotlib.pyplot as plt

Below you import both matplotlib and numpy to create a plot.

import matplotlib.pyplot as plt
import numpy as np

# Generate some sample data
x = np.linspace(0, 2 * np.pi, 100)
y = np.sin(x)

# Create a prettier plot
plt.figure(figsize=(8, 6))  # Set the figure size
plt.plot(
    x,
    y,
    label="Sine Wave",
    color="b",
    linewidth=2,
    linestyle="-",
    marker="o",
    markersize=6,
    markerfacecolor="r",
)
plt.title("Prettier Matplotlib Plot")
plt.xlabel("X-axis")
plt.ylabel("Y-axis")
plt.grid(True, linestyle="--", alpha=0.7)
plt.legend(loc="upper right")
plt.tight_layout()

# Show the plot
plt.show()