Install a new kernel in Jupyter Notebook using a specific Python environment

Summary

Instead of running a separate instance of Jupyter Notebook for different Python environments, it is possible to install a kernel with a specific Python environment in Jupyter Notebook. The environment is then configured when creating a new notebook.

Procedure

To install a kernel with a specific Python environment in Jupyter Notebook, follow the steps described below:

1. Run the Python Command Prompt as an administrator.

Note:
The Python Command Prompt window opens with the active default Python environment.

2. In the Python Command Prompt window, insert the following command:

python -m ipykernel install --user --name <python environment> --display-name "Python (<python environment>)"

3. Press Enter on the keyboard, and the following response is returned:

Installed kernelspec <python environment> in C:\Users\<user>\AppData\Roaming\jupyter\kernels\<python environment>

The image below shows the command and response in the Python Command Prompt window. The name of the Python environment used in this example is arcgispro-py3-clone1.

4. Run Jupyter Notebook from the default Python environment.
5. In the Files tab, click New.

Alternatively, manually create a new kernel with a specific Python environment in the kernels folder.

1. Navigate to the following location:

C:\Users\<Username>\AppData\Roaming\jupyter\kernels

2. Create a kernel JSON file.
   1. Using Notepad++, insert the following JSON code:

{
"argv": [
"C:\Users\<user>\AppData\Local\ESRI\conda\envs\<python environment>\python.exe",
"-m",
"ipykernel_launcher",
"-f",
"{connection_file}"
],
"display_name": "Python (<python environment>)",
"language": "python"
}

2. In the Notepad++ window, click File > Save As... .
3. In the box next to File name, insert kernel.json.
4. Click the box next to Save as type, and select All Files.
5. Click Save.