To kick-start with data science and machine learning, most important is environment setup. Jupyter Lab is next-generation platform similar to Jupyter Notebook. We will install Anaconda on Ubuntu16.04 VPS server, which will contain an all-new version of Jupyter Lab.
What is Anaconda?
Anaconda is python distribution for data science and machine learning. It has a collection of 250+ different packages that are widely used for data science and machine learning. It supports most of operating systems like Windows, Linux and Mac.
Jupyter Lab Features
- Next-generation user interface for Jupyter Notebooks
- Based on Jupyter Notebook and Architecture
- Most important, it has a console support. So you manage your VPS server without ssh login.
Let’s install Jupyter Lab on Ubuntu 16.04. Login to your VPS server using ssh client and follow these steps:
Download Anancode script from the following link. Please check the latest version here and update link accordingly.
Run following commands
During installation, it will ask for environment details. If you are aware of it, you can modify details or proceed with default one. If you have selected to add conda to Linux bash, run following command to update.
Before we run Jupyter Lab, it’s necessary to protect it with a password. If you don’t set a password, then you will need authorization token. This token will change every time when you run notebook. For convenience, it’s better to generate a password.
jupyter notebook password
We are ready to run our Jupyter Lab.
jupyter lab --port 9999 --no-browser
Log in with the password and your platform is ready to start with machine learning and data science. Please keep in following things.
- We should secure our Jupyter Notebook/Lab with SSL and other security measures. Incoming article I’ll cover how we can secure our environment for public servers.
- By default, we are using root environment, which is not advisable. It’s always better to create the custom environment. Most probably I’ll cover that also.
- Since we have installed Anaconda, still you can use old Jupyter Notebook with the following command.
jupyter notebook --port 9999 --no-browser