Skip to content

This project represents a whole process of Anime data collection, preparation, and delivery as a data app, powered by technologies like Pandas, Jupyter, and Streamlit

Notifications You must be signed in to change notification settings

lmassaoy/myanimelist-data-collector

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 

Repository files navigation

myanimelist-data-collector

Special thanks to this AMAZING website! blink-emoji-README.png

mal-logo-README.png

Intro

This project relies on the famous and trusted online anime database myanimelist.net. It's a complete, well-architected and easy-to-use website that allows us, fans of animes/mangas, to navigate through our favorite Asian arts.

Its community is huge and engaged, participating in sharing reviews, applying ratings, and discussing the tv shows and HQs.

Trusting in the website and its community/forum, we manage to scrape the data of anime/manga profiles to build a rich dataset to analyze, tell stories and recommend new series for you to watch, based on your favorite styles and series you've watched before.

Take a look into the macro view of the project!

solutions-macro-view.png

Architecture

architecture-README.png

ETL

Extraction: Web scrap job

Important: Remember to interval the extractions to do not badly influence other users' experiences while navigating the website. Thank you! blink-emoji-README.png

Run it as a normal python job, passing the start and end 'anime ID' as arguments:

$ python3 code/python/data-collector.py -s 1 -e 50000

This job will collect the data from HTML through web scrap using the library BeautifulSoup4 and persist as json files in the raw data layer.

Transform and Load job

Run it as a normal python job

$ python3 code/python/data-transformer.py

This job will read all json files, transforming in a single dataframe using Pandas, enhancing the data and finally saving it as a parquet file in the enhanced data layer.

Services (Data App and Sandbox)

We're gonna light them up by using a Docker stack. As simple as this:

  1. build the image for the Data App
$ myanimelist-data-collector % docker build -t anime-data-app:latest -f code/data-app/Dockerfile .
  1. once the image is available, we can execute our two containers with Docker compose:
$ myanimelist-data-collector % cd devops
$ devops % docker-compose up -d
Creating network "jupyterNetwork" with driver "bridge"
Creating anime-app ... done
Creating jupy      ... done

Data App (BI)

Access: http://localhost:10001/

welcome-data-app.png

As you can see in the left panel, you can turn on the features and apply filters to the datasets. For example:

sample-data-app.png

The concept of the features are:

  • KPI: core metrics
  • TOP N: exploring the data by ordering them and selecting the most relevant
  • Data Visualizations: visual data analysis
  • X-Ray: deep dive into a single anime title

Sandbox (Ad-hoc)

Access: http://localhost:10000/

login-jupy.jpeg

Oops, you need to login here. Let's get this opportunity to install some dependencies too :)

Access the terminal of the container 'jupy' and install some requirements. Once its done, check its log and get the token to access the Jupyter.

turning-on-jupy.jpeg

Now you're good to go blink-emoji-README.png

jupy-logged.jpeg

  1. to turn off the containers, you can use this:
$ devops % docker-compose down
Stopping anime-app ... done
Stopping jupy      ... done
Removing anime-app ... done
Removing jupy      ... done
Removing network jupyterNetwork

About

This project represents a whole process of Anime data collection, preparation, and delivery as a data app, powered by technologies like Pandas, Jupyter, and Streamlit

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published