Tools for Data Science - Jupyter Notebooks and JupyterLab

Tools for Data Science

Jupyter Notebooks and Jupyter Lab

Introduction to Jupyter Notebooks

Origin of Jupyter Notebook

- Jupyter Notbeooks : Originated as "iPython", stands for JUlia PYThon R

 

What is a Jupyter Notebook?

- browser-based application

- allows to create and share documents containing code, equations, visualizations, narrative text links, ...

- records data science experiments

- allows combining text, code blocks, and code output in a single file

- exports the notebook to a PDF or HTML file format

 

JupyterLab

- allows access to multiple Jupyter Notebook files, other code, and data files

- enables working in an integrated manner

- is compatible with serveral file formats

- is an open source

 

Getting Started with Jupyter

Recap

- run, delete, and insert a code cell

- run multiple notebooks at the same time

- present a notebook using a combination of Markdown and code cells

- shut down your notebook sessions after you have completed your work on them

 

**

그냥 이래저래... jupyterlab 어떻게 쓰는지 클릭하면서 보여줬다

근데 나 이정도는 할 수 있는 사람이걸랑...!! 싶어서 대충 보고 넘겼당... 헤헤

 

근데 이 다음에 나오는 hands-on-lab 되게 좋다!!

launch app 나오면 바로 jupyter lab 토이 사이트?가 나온다

당연히 앱 다운로드일 줄 알았는데 ㅎㅎ 신기하고 좋다

**

Jupyter Kernels

What is a kernel?

- is a computational engine that executes the code contained in a notebook file

- exists for many languages

- launches when a jupyter notebook is opened

- performs the computation when the notebook is executed

- can switch to a different kernel as per your requirement

 

** 

여기서도 나오는 launch app!

강의 대충 들어서 다 solution 어떻게 되어있나 확인하긴 했지만... ㅎㅎ

넘 좋다 실습까지 보장되어있는 코세라!! 최고최고!! (광고 아닙니다)

**

 

Jupyter Architecture

Architecture

- represents code, metadata, contents, and outputs

- user -> browser -> notebook server -> notebook file / kernel

- jupyter implements a two-process model with a kernel and a client

- the notebook server is responsible for saving and loading the notebooks

- kernel executes the cells of code contained in the notebook

- the jupyter architecture uses the NB convert tool to convert files to other formats

 

Additional Anaconda Jupyter Environments

Computational Notebooks

- combine code, compuational output, explanatory text, and multimedia resources in a single document

 

JupyterLab

- an open-source, web-based application

- enables the creation of code, interactive visualizations, text, and equations

- includes pre-installed Python libraries (eg. NumPy, Pandas, Matplotlib)

 

Anaconda

- a free and open-source distributor for python and R

- has 1500+ libraries

- free to install

- free community support

- installs new packages without a CLI

 

VisualStudio Code (VSCode)

- a free, open-source code editor for operations like debugging and task running

- works on Linux, Windows, and MacOS

- supports : multiple languages, syntax highlighting, auto-indentation

- one of the most popular devlopment environment tools

 

Additional Cloud Based Jupyter Environments

JupyterLite

- lightweight tool built from JupyterLab components

- executes in the driver

- dedicated Jupyter server not required

- can deploy as a static website

- can create interactive graphics and visualizations

- supports visualization libraries like Altair, Plotly, and ipywidgets

- includes JupyterLab's latest improvements and features

 

Google Colaboratory (Google Colab)

- free Jupyter notebook environment that runs entirely in the cloud

- execute on a browser

- store on Google Drive and Github

- upload and share without setup and installation

- clone from Github and execute in GoogleColab

- most libaries are pre-installed (scikit-learn, matplotlib)

 

Practice Quiz - Jupyter Notebooks and Jupyter Lab

Q. What functionality is provided by the Jupyter Notebook file to Data Scientists?

A. Allows to combine descriptive text, code blocks, and code output in a single file

 

Q. What is the purpose of the kernel in the Jupyter Notebook?

A. Executes the code

 

Q. What are the elements of Jupyter's architecture?

A. Kernel and client

 

Q. What is the default kernel of JupyterLite?

A. Pyolite

 

Q. What is the purpose of Anaconda?

A. Allows to install new packages without using a command line interface