Docker Image for Environmental Data Science

Author: Dr. Zhonghua Zheng (zhonghua.zheng@outlook.com)

Introduction

This Docker image is used for Environmental Data Science. It consists of:

• Jupyter
• data processing
  • numpy
  • pandas
  • xarray (with netCDF4 support)
  • scipy
• machine learning
  • scikit-learn
  • PyTorch (for Deep Learning)
  • LightGBM
  • XGBoost
• mapping & plotting:
  • Cartopy
  • matplotlib

Usage

Step 0: Install Docker Desktop. It's free.

• Download from here.

Step 1: run a docker container (using Terminal)

• If you have your scripts (notebooks) and data in the same folder:

  $ cd your_scripts_folder
  $ docker run -it --rm -p 8888:8888 -v $PWD:/home envdes/env

• If you have your scripts (notebooks) and data in different folders (then you would have a folder "/data" after you running the container):

  $ cd your_scripts_folder
  $ docker run -it --rm -p 8888:8888 -v $PWD:/home -v /path_to_data_folder:/data envdes/env

Step 2: open your browser, type "localhost:8888", copy the token from terminal (after "token=") and paste into the webpage

• When you import python packge(s), if you got the error like:

  libgomp-d22c30c5.so.1.0.0: cannot allocate memory in static TLS block'
  

  please import the package(s) based on the order below

  import sklearn
  import torch
  import xgboost
  import xarray as xr