Geospatial Analysis Accelerator

The objective of this repository is to:

Identify, demonstrate and create assets that can be used by Data Scientists to speed up their geospatial analysis.

These assets are the result of an research initiative started by the IBM Academy of Technology. The intend is to extend and grow the number & maturity of assets over time. This repository will hold the assets supporting that initiative.

The background of the IBM AoT Study and the first Use Case on Spearman are described in two LinkedIn articles:

• Accelerate analyzing the influence of regional factors on COVID-19 using IBM PAIRS & Jupyter Notebooks
• Spearman correlations between the UV Index and COVID-19 per region using IBM PAIRS & Jupyter Notebooks

An overview of the first asset is also published on LinkedIn, with a more detailed description in the Notebook itself:

• Geospatial Analysis Accelerator for COVID-19 & IBM PAIRS

The second version of the asset, which includes support for a generalized additive model (GAM) and the inclusion of R, is described on LinkedIn as well:

• GAM associations between regional weather factors and COVID-19 using IBM PAIRS, Jupyter Notebooks and R

COVID-19 Geospatial Analysis

The first asset is a Jupyter Notebook that can be used to determine Spearman's rank correlation coefficient or a log-linear generalized additive model (GAM) between COVID-19 cases and Geospatial & Temporal information out of IBM PAIRS Geoscope.

The COVID-19 Geospatial Correlation Notebook can be used in a existing Juyter Notebook server, provided that Python 3.x is used and the required packages are installed via pip and conda.

The second option is to use the provided Dockerfile to build a Docker image. To use this option you require a Docker environment, e.g. by installing Docker Desktop.

Then you need to build the Docker Image:

1. Clone this repository and use a terminal to navigate to the cloned directory.

2. Build the Docker Image:

   docker build --rm -t ibm/geospatial-analysis-accelerator .
   

3. Start the Docker Image & Mount the current directory:

   On Linux & Mac:

   docker run -it --rm -v $(pwd):/home/jovyan/work -p 8888:8888 ibm/geospatial-analysis-accelerator
   

   On Windows via cmd:

   docker run -it --rm -v "%cd%":/home/jovyan/work -p 8888:8888 ibm/geospatial-analysis-accelerator
   

   On Windows via Powershell:

   docker run -it --rm -v ${pwd}:/home/jovyan/work -p 8888:8888 ibm/geospatial-analysis-accelerator
   

   This will start the Jupyter Notebook server and mount the current directory.

   The Jupyter Notebook server will be available on http://127.0.0.1:8888/ with the access token listed in the console.

The Notebook is available in the notebooks directory and contains full documentation on how to use the Notebook.

IBM PAIRS

If you don't have access to the IBM PAIRS API, then you can request accesss via the IBM PAIRS Access procedure.

If you want to get started without API access you can ! This repo contains an IBM PAIRS Local Cache file for The Netherlands.

The file contains a small set of data:

• geospatial filter: NL
• temporal filter: February 13 2020 - June 19 2020
• data filter: IBM TWC (UV Index, Temperature Above Ground, Relative Humidity, Wind Speed).
• aggregation: daily by region code

This small set of data is sufficient to run the COVID-19 Geospatial Correlation Notebook, but true value comes with access to the IBM PAIRS API.

The data in this IBM PAIRS Local Cache File is made available under CDLA-Sharing-1.0

Disclaimer

The exploration is only to identify & create assets for data scientists to explore geospatial-temporal data. Examples provided with the assets should not be taken as any interpretation of the results. We are not trained epidemiologists and therefore leave all interpretations to those that have the professional expertise.