https://github.com/sodascience/kansenkaart_analysis

Analysis code for the kansenkaart project
https://github.com/sodascience/kansenkaart_analysis

Last synced: about 1 year ago
JSON representation

Analysis code for the kansenkaart project

• Host: GitHub
• URL: https://github.com/sodascience/kansenkaart_analysis
• Owner: sodascience
• License: mit
• Created: 2021-06-09T14:25:02.000Z (about 5 years ago)
• Default Branch: main
• Last Pushed: 2024-10-31T11:01:19.000Z (over 1 year ago)
• Last Synced: 2025-04-12T08:13:20.506Z (about 1 year ago)
• Language: R
• Size: 175 KB
• Stars: 2
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: readme.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Parallel estimation for the Kansenkaart on the OSSC

![pipeline.png](pipeline.png)

This repository holds the code for parallel estimation of the Kansenkaart project estimates on the ODISSEI Secure Supercomputer.

This is a work in progress!

## Steps to run the analysis

### Repository preparation

1. Add the cohort datasets resulting from the [`preprocessing pipeline`]() to the `data/` folder.

2. Point to the right resources in lines 8-9 in `02_create_grid.R`.

3. Ensure the right COROP code year (default: 2019) is being used in line 72 of `02_create_grid.R`.

4. Check that the right datasets are loaded in lines 33-37 of `03_array_job.R`.

### OSSC running

1. Create and store the `data/model_grid.feather` file by running `02_create_grid.R`. 

1. Perform a test run by running `Rscript 03_array_job.R 1` in bash on the OSSC

3. Check how many jobs are needed (`nrow(model_grid) / batch_size`)

2. Create an array job using `array_job.sh`.