Using Participatory Input to Navigate the Machine Learning Multiverse

This document is an interactive companion article illustrating a subset of the results from the paper "Preventing Harmful Data Practices by using Participatory Input to Navigate the Machine Learning Multiverse".

For a detailed explanation of the methodology, analysis and decision paths explored, please refer to the Methods and Results sections of the paper.

This interactive analysis allows for the exploration of the participatory multiverse for individual countries. You can select a country from the dropdown below to view its particular figure.

After selecting a country, you can click individual nodes in the analysis to see the actual path that leads to a node. The further to the right of the figure you select a node, the longer, thinner and more detailed a path will be. Hovering over a node will display a tooltip with the decision and choice that node represents.

To protect participant privacy, we only generate separate figures for countries with at least N >= 5 participants after filtering. To reduce the complexity of the figures we only include data from participants with complete data for all decisions for this analysis. The full list of counts per country is available in Table 1, as well as information on whether or not this country's data is visible in a separate plot or in aggregate.

Table 1. Number of participants per country included in the figure below. To avoid further complexity, only data from participants with complete responses for all decisions were included for these figures.