Many scientists base their data-analysis on scripts, i.e. Python based analysis scripts. Science is deploying a Jupyter based data analysis processing farm, and it would be great for scientists to be able to use task parallelism for their data analysis. This should work in a way where a user makes an analysis script work on a dataset, and once it works the same script should be applied to all parts of a dataset, i.e. once analysis of one image works the same analysis should be done on the remaining 1000 images.