HowToCodePage

Installing JupyterLab and managing kernels

Notebook, JupyterLab and Notebook 7 compared, installing into a virtual environment, and why a kernel is not the same thing as an interpreter.

Three front ends, one protocol

Every Jupyter front end talks the same kernel protocol: the browser sends code, a separate process executes it and returns outputs. Which interface you install only changes the editor, not how code runs.

Front endPackageBest for
JupyterLabjupyterlabDaily work: tabs, split panes, terminal, file browser, extensions
Notebook 7notebookThe classic single-document UI, now built on the same components as Lab
VS Code / editorsbuilt inNotebooks next to the rest of your codebase and debugger
nbclassicnbclassicLegacy extensions only; rarely worth it on a new machine

JupyterLab can open classic .ipynb files, so choosing Lab does not lock you out of anything.

Install into the environment that has your libraries

python -m venv .venv
source .venv/bin/activate          # Windows: .venv\Scripts\activate

pip install jupyterlab ipykernel numpy pandas
jupyter lab

Installing Jupyter globally and your libraries in a venv is the most common way to end up with a kernel that cannot import anything you expected.

💡
A kernel is a separate process launched by Jupyter, not a thread inside the server. It inherits the interpreter that has ipykernel installed, which is why ipykernel must live in each environment you want to use.

Registering and listing kernels

# register the active environment under a readable name
python -m ipykernel install --user --name sales-analysis --display-name "Python (sales)"

jupyter kernelspec list              # where each kernel points
jupyter kernelspec remove sales-analysis
# verify inside a notebook which interpreter is really running
import sys
print(sys.executable)
print(sys.version)

# and where it will look for packages
print(sys.path[:3])
  • The kernel name is the identifier used on the command line; the display name is what the menu shows.
  • --user writes to your home directory; drop it to install for all users of that Python.
  • In Lab, switch kernel from the kernel selector in the top-right of the notebook.

FAQ

The notebook cannot import pandas but my terminal can. Why?
You are running a different interpreter. Print sys.executable in a cell and compare it with which python. Register a kernel from the environment that has your packages installed.
Should I use conda or venv?
Either, as long as you stay consistent. conda is convenient for binary scientific packages; pip plus venv is lighter and closer to what production uses. Do not mix both managers inside one environment.

Notebook fundamentals Magics: line, cell and shell commands

Last refreshed 2026-09-18.