Can machine help me decide the invitation list for next Friday?

Click on image to view the interactive visualisation

Who to invite next Friday?

This is exploration for understanding machine learning models and their decision making process works.

As an interaction designer, I wondered can machine learning algorithms, understand my social life and suggest whom to invite for the coming Friday for dinner and drinks?

Is it possible for machines (ML models) to predict and suggest possible and feasible choices of social groups that one can trust to act on? If all possible informations and parameters are provided like - what is valuable, what parameters is important, who is important, why they are important, can the machines then equate and objectify such complex human connections?
Can then machines equate relatively feasible choice of friends to go out with on Friday night?

This project used my personal Facebook data and mobile call data, call history, chat history to come back with possible answers to whom I should invite.

IF it is possible to understand such complex human connections and social relationships through simple equations. The the question is how will it show the answers, so that users could analyse and understand the logics and also to some extent provide a feedback regarding the decisions

The project's aim was to allow us to understand what all parameters an algorithm is taking into consideration to come back with the right answer. This was our attempt to bring transparency to machine learning algorithms and allowing users a small chance to manipulate their own data.

Critically looking at the project, whether allowing users to manipulate and peek inside the Machine learning algorithms are the question that we need to be addressed next.

More details

In collaboration with Microsoft Research, Cambridge.
Supervised by David Sweeney, Human Experience and Design Group.
Creative coding done in collaboration with Michael Kalygin using D3.js

More links . ︎  ︎  ︎