Convenient new features of Jupyter Lab Conclusion

An environment construction tool (IDE) “Jupyter Notebook” that machinery learning and data science engineers love so much. This time, the alpha version of Jupyer Lab (Jupiter / Laboratory) which was released before was officially released as a beta version again !

Even codexa teams have many members who transferred from traditional notebooks to Jupyter Lab. In this article, we have summarized usability, merit / disadvantage, etc. of a newly added wonderful new function about IDE’s decided version of machine learning “Jupyter Lab”. It is also a recommended development environment tool for beginners who are planning to learn machine learning from now, so please refer.

Machine learning introductory tutorial using Jupyter

It is unnecessary to build Jupyter’s environment, it is open to the machine learning introductory tutorial which can be executed online! Why do not you jump into the machine learning world by first learning basic algorithms?

  • Coding the least square method and the steepest descent method with Python with scratch (linear regression)
  • Outline of logistic regression and knowledge useful for mathematical understanding and practice (logistic regression)

What is Jupyter Lab? About the Jupyter project

First of all is the Jupyter project? about. Jupyter (Read Jupiter) is a project to develop open source interactive computing (Interactive Computing).

This Jupyter project developed Jupyter Notebook. Jupyter Notebook is a coding environment that can be used on a browser, allowing you to share code with colleagues and teams, develop interactive analysis results, and easily integrate and handle large-scale data. It is one of the most popular development environments (IDE) that machinery learning engineers around the world use it because of its ease and ease of use.

Although I have been using it for many years, Jupyter Notebook can illustrate the entire process of data analysis in a simple and clean form, and I think that there is a great merit to further sharing.

Although it is not easy for each team to write code or analyze data, it is not easy to integrate code in each cell in Jupyter Notebook, and each time the output (output result) Because it accompanies it, it is possible to understand even a code written by another person in a very short time with little effort .

As Jupyter Lab introduced this time, development project is proceeding as an IDE to make it possible to perform data scientist and machine learning engineer’s work process integrally and efficiently.

If you are already an active engineer, I think that some IDE (Integrated Development Environment) is used. Very simple, Jupyter Lab will be an IDE for machine learning and data science .

Although it is Jupyter Lab, it is very similar to the conventional Juyter Notebook, but many new features have been added! In this article, I have summarized how to install Jupyter Lab, even merits and demerits!

Installation of Jupyter Lab

Well, first, I will explain how to install Jupyter Lab. Since it is installable on Pip or Anaconda, it is easy to install with either one if you can use either environment.

pip install jupyterlab
It is a more detailed installation method, but it is gathered in the official document (English) . I have already installed on 4 machines, but I did not get any errors. If you get an error please see the official document.

Launch Jupyter Lab

After successful installation, let’s start Jupyter Lab immediately. If you are a Mac user terminal, if you are using Anaconda on Windows start up with Anaconda with the following command.

jupyter lab

When Jupyter Lab starts up, it automatically opens the Jupyter Lab interface with the URL “http: // localhost: 8888 / lab” in the default browser. If you are using a conventional Jupyter Notebook, it will be exciting for new features at this timing (tab !!!).

Advantages / Disadvantages of Jupyter Lab

I actually used Jupyter Lab in several projects, so I would like to briefly summarize the merits and disadvantages that individuals think. (It is my personal opinion to the last)


  • A more sophisticated interface that touches with a sense similar to a notebook
  • tab! (One press)
  • Multiple window display such as notebook, text, CSV, console is possible
  • Integration with Google Drive
  • Cells that can be dragged and dropped
  • File explorer (one press)


  • If you have the ability to edit variables like R Studio IDE, ◎
  • Doing heavy processing will freeze

Basically there are few disadvantages, so if you are familiar with Jupyter Notebook, I think that it is better to transfer to Jupyter Lab ! However, there are also parts where the action (?) Still seems unstable, and there are cases where JupyterLab stopped when training a huge amount of data etc. (Notebook was completed without problems with exactly equivalent code).

New features of Jupyter Lab

Well, from here on we will introduce some useful new features added to Jupyter Lab.

New innovated interface

Although it briefly mentioned the above merit already, I think that it is the interface renewal that noticed at the very first by transferring from Notebook to Jupyter Lab. Although it is an interface of a conventional Notebook, it was very simple, I liked it, but Jupyter Lab added a number of new functions with high practicality in addition to simple.

As a new function added to the interface, a file viewer was installed (finally!). It is also useful to check the currently running kernel from the sidebar. With this side bar and the newly added “Tab” function, Jupyter Noteobok made moving between multiple kernels very smooth.

As you can see from the capture below, almost all interfaces have been improved, but it is also a pleasing point for notebooks to be usable with ease of use ! The function that I wanted is attached, furthermore, the impression that the function which I do not want to change is intact as it is.

Display data file in table format

It is a new function added to Jupyter Lab, but it is a new function that displays data such as CSV file in table format . Honesty · This is really useful. It is a function attached with R Studio (R language IDE), but it is saved by becoming usable also with Jupyter.

In machine learning, confirming data is quite frequent work, so it is easy to check files on Jupyter easily as this way shortens work time.

Multiple window (multiple window display)

This is also one of the useful new features. The first time I transferred from Jupyter Notebook was awkward to use, but it is a function that I can not relinquish if I get used to it.

In conventional notebook, if there are multiple files to use, it was necessary to open multiple windows. By dragging and dropping multiple windows with this new function, it is possible to place it in a favorite place. Especially the work of coding while checking the data set became dramatically easier.

As in the capture below, while coding on the main screen Notebook, you can check the data set in the adjoining auxiliary window and work out with the console further down.

Can work with Google Drive

I think that there are an increasing number of cases using Google Drive in offices and teams, but exactly this new feature is perfect for that kind of people.

Although this new feature is installed, it becomes possible to use it by installing the extension function and logging in to the Google account from the side menu of Jupyter Lab. What you can do is save the notebook you made in Jupyter Lab directly to Google Drive and even share it inside the team.

Each time you save the notebook and send it by e-mail, it will be gone. If you have a team that already uses Google Drive, let’s set it up!

Cell drag & drop

Well, in transferring to Jupyter Lab this time, if you attach a convenience ranking of new functions … · Personally this function becomes first prize first … · · · That is cell drag & drop !

If you are already using Jupyter Notebook, perhaps you have not felt at least once? Vertical movement of the cell and place movement is very troublesome. It was very troublesome to edit the order of the code etc, such as writing the code properly at the beginning and adding it later.

This new function solves such useless work! It is hard to understand with a slightly lower capture, but you can grab a designated cell and drop it to a favorite place.


How was it? This time I introduced new useful functions of Jupyter Lab officially released as a beta version, and also summarized merits / demerits.

Although it is Jupyter Lab, I think that it is a tool that is used by a very wide range of layers from skilled machine learning engineers to first scholars. It is still under development as a beta version. I’m looking forward to how it will change from now on! As a personal request · · I think that it will be nice if variable viewer is added like R Studio IDE!

Even beginners who begin machine learning from now on, machine learning engineers who are already stirring data in Gorigori, let’s try using Jupyter Lab by all means! As machine learning environment (IDE), you probably will not find any more options if you look for it.


3 thoughts on “Convenient new features of Jupyter Lab Conclusion

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.