Introduction

The Data Scientist Association has released ** "Data Science 100 Knock (Structured Data Processing)" **, a free learning environment where you can practically learn how to process structured data [on GitHub](https: / /github.com/The-Japan-DataScientist-Society/100knocks-preprocess). This article describes the details of the introduction procedure so that even beginners can build a free learning environment. (The execution environment to be built is shown in the figure below.)

Prerequisites (Windows10)

Docker Desktop for Windows

If it does not start, Hyper-V may be "disabled", so set it to "enabled". ](Https://docs.microsoft.com/en-us/virtualization/hyper-v-on-windows/quick-start/enable-hyper-v#enable-the-hyper-v-role-through-settings)

Git for Windows

Set the line feed code change, which is the default setting at the time of installation, to "disabled".
The commands in this article have been executed and verified with PowerShell.

> git config --global core.autocrlf input

Environment

Create a directory for the learning environment (dss this time) and clone a repository of 100 knocks. After that, move to the 100 knock directory and use the docker-compose command to create a container. (It takes about 10 minutes.)

If a warning pop-up is displayed while building the environment, you may not have access authority to Docker's local PC, so select ** "Share it" ** to grant access authority.

> mkdir dss
> cd dss
> git clone https://github.com/The-Japan-DataScientist-Society/100knocks-preprocess.git
> cd 100knocks-preprocess
> docker-compose up -d --build

If you can check the started container and check the output of ** "dss-notebook" ** and ** "dss-postgres" **, the environment construction is successful.

> docker ps

CONTAINER ID        IMAGE               COMMAND                  CREATED             STATUS              PORTS                    NAMES
b35f99d4148a        dss-notebook        "tini -g -- start-no…"   23 seconds ago      Up 22 seconds       0.0.0.0:8888->8888/tcp   dss-notebook
3cb559c7f66d        dss-postgres        "docker-entrypoint.s…"   27 seconds ago      Up 26 seconds       0.0.0.0:5432->5432/tcp   dss-postgres

How to use

You can access the built Jupyter environment by accessing the following URL with a browser.

http://localhost:8888

Under the work directory, there is an .ipynb file for structured data processing exercises. ** Import of required library and data acquisition before processing are already described in the first cell. ** ** Enter the process suitable for the exercise in the blank cell and execute it to proceed with the learning. The answer to the exercise is in the .ipynb file in the work / answer directory. Therefore, you can work while checking the correctness of the processing answered in the exercise file.

Stop / start learning environment

You can stop the built environment with the following command.

> docker-compose stop

Also, if you want to start it after the second time, you can start it with the following command.

> docker-compose start

Supplementary information

When the response of the built environment is slow

Change the Memory value of Resources in Settings of Docker Desktop for Windows. The recommendation is 4.00GB or more.

If port 8888 is in use

If you are using the 8888 port of the local host in another development environment (LAMP etc.), you can handle it by changing docker-compose.yml as follows (change the port value of notebook).

`docker-compose.yml`


  notebook:
    ports:
      - "888:8888"

In the above case, it will be accessible at the following URL.

http://localhost:888

Summary

Described the environment construction procedure for 100 data science knocks (structured data processing) in the Windows 10 environment. If you have any questions or concerns regarding the above procedure, we would appreciate it if you could comment.

Reference link

Data Science 100 Knock Guide

[PYTHON] Data science 100 knock (structured data processing) environment construction (Windows10)