How to download kaggle dataset?

Question

How do I download kaggle datasets to colab or any other place from script or notebook?

! kaggle datasets download -d arslanali4343/world-cities-database-population-oct2022

Answer 1

You can try this one using Colab

Step 1:

from google.colab import files

files.upload() # <<-- Upload your Kaggle API-Token 

! mkdir ~/.kaggle # <<-- Create a Kaggle folder inside Colab environment 

! cp kaggle.json ~/.kaggle/  # <<- Gets your Token into it

! chmod 600 ~/.kaggle/kaggle.json  # <<- Gives Colab permissions

! kaggle competitions download -c <competition_name>

file_path = '/content/<competition_file_name>'

After that, you'll get a .zip file, so...

Step 2:

from zipfile import ZipFile

with ZipFile(file_path, 'r') as unzip:

    unzip.printdir() # <<-- Check inside zip file

    unzip.extract('train.csv') # <<-- Extract a single file

    #unzip.extractall() # <<-- Extract all files at once
    

You can also use a submission command to upload your files straight to Kaggle servers

! kaggle competitions submit <competition_name> -f <submission_file> -m "My submission message"

Have fun