How to fix ModuleNotFoundError in Jupyter Notebook & Google Colab

ModuleNotFoundError is one of the first “real” errors most Python and ML students hit in Jupyter or Google Colab. You install something, import it, and boom:

ModuleNotFoundError: No module named 'sklearn'

The good news: it’s almost always fixable with a few checks. In this guide, we’ll walk through a simple mental model to **diagnose → fix → prevent** ModuleNotFoundError – and why it matters for keeping your notebooks reproducible.

Step 1: Understand what ModuleNotFoundError really means

Python is telling you: “I looked in all the places I know about for this package, and it’s not there.” That could be because:

• the package isn’t installed in this environment,
• you installed it in a different environment (or different Python version),
• you misspelled the module name,
• you’re trying to import a local file that isn’t on the Python path.

Step 2: Fix ModuleNotFoundError in Jupyter Notebook (local)

1. Check which Python you’re using inside the notebook

In a cell, run:

import sys
print(sys.executable)
print(sys.version)

This tells you which Python binary your notebook is using. Make sure you installed the package into this environment, not some other one.

2. Install the package from inside the notebook

Best pattern: install using pip or conda directly in the notebook:

!pip install scikit-learn

import sklearn

If you’re using Conda and Jupyter together, check that the notebook kernel is tied to the correct Conda environment where the package exists.

3. Fix imports for local modules

If you see something like:

ModuleNotFoundError: No module named 'config'

but you have config.py in your project, check your folder structure:

project/
  notebook.ipynb
  config.py

If the notebook is in a subfolder (e.g. notebooks/notebook.ipynb), you may need to adjust sys.path or move the file so Python can find it.

Step 3: Fix ModuleNotFoundError in Google Colab

1. Install packages in the Colab runtime

Installing Python packages on your laptop doesn’t affect Colab. It runs on Google’s servers, in a fresh environment each time.

Install from a cell at the top of your notebook:

!pip install scikit-learn pandas matplotlib

2. Re-run installs after runtime reset

When Colab disconnects or the runtime resets, installed packages vanish. If you see ModuleNotFoundError again after a reset, just re-run your installation cell.

3. Handle custom / private packages

If you’ve got a private package on GitHub or a local module, use:

!pip install git+https://github.com/username/repo.git

or mount Google Drive and adjust the Python path so Colab can see your module.

Step 4: Make your fix reproducible (for future-you and others)

Fixing the error once is good. Making sure you never have to debug it again on this project is better. That means:

• keeping a list of your dependencies, and
• capturing the project state when everything works.

1. Capture dependencies in a file

From your notebook environment, run:

!pip freeze > requirements.txt

Commit this file to your project or at least keep it next to your notebook. Anyone wanting to run your project can install from it:

pip install -r requirements.txt

2. Use NoteCapsule to snapshot code + deps + data layout

To go one level beyond just fixing the error, you can use NoteCapsule to create a Capsule:

!pip install notebookcapsule -q

from notebookcapsule import create_capsule

create_capsule(
    name="after-fixing-modulenotfounderror",
    data_dirs=["./data"]
)

This writes a folder under ./capsules/ with:

• your notebook,
• requirements_suggested.txt (from your imports),
• data_manifest.json listing the files you use,
• a README_template.md to explain how to run the project.

That way, the next time you (or someone else) run this project, you’re far less likely to hit ModuleNotFoundError again.

Quick checklist for fixing ModuleNotFoundError