Detailed Results and Visualizations

Detailed results is a means of passing extra information from the scoring program to the frontend.

This is done via writing to a detailed_results.html file (OR any .html -- first by alphabetical order -- in the output folder), and setting enable_detailed_results to True in competition settings (via yaml or editor).

This file is watched for changes and updated on the frontend every time the file is updated, so users can get a live feed from the compute worker.

There is no limitation to the contents of this HTML file, and can thus be used to relay any information desired. Use case ideas:

• Plot data using a python plot library like matplotlib or seaborn.
  • plot the learning curve over time of a reinforcement learning challenge
  • plot the slope of a linear regression model
  • plot the location of clusters in a classification challenge
  • plot anything you can conceive of
• Run a profiler that outputs a network of method calls.
• Display any additional data about the submission file that can not be distilled down in to a score of some kind