Behind the scenes
↓

Behind the scenes

Purpose

The purpose of 'The Oracle' project is to demonstrate how my classmates' data can be used to manipulate their emotions and behaviours. My algorithm extracts insights from their data, and then turns these insights into social 'nudges' by comparing them to other people. This may prompt a positive or negative emotional response, or influence future behaviour.

They may be surprised, pleased, annoyed or upset about the conclusions that their data provides – but ultimately, they will have to confront the fact that their own data is the source of this judgement.

The Oracle allows users to provide emotional feedback

Post-corona context

Coronavirus has provided an unprecedented justification for data harvesting: people are far more willing to surrender their data if they believe that their health is at stake. This experience intends to prompt my classmates to view their personal data in a wider social context, and reconsider whether or not their health is worth sacrificing their privacy and freedom for in a post-corona society.

This moment in history is important for determining the future of our relationship with data and social control. This is an opportunity to set a precedent: how much data can a goverment harvest from its population and still get away with it, under the guise of 'safeguarding public health'? How much control will we, as a population, allow our governments to have over our behaviour?

Are we able to tell when we are being manipulated?

Are we willing to allow it for the good of society at large?

The algorithms

Describing and planning potential algorithms

1: Purpose as proportion of total

2: Narcissism as proportion of total

1/3

Example:
"Your dataset is the most narcissistic in the class."

2/3

3: Earliest data entry

Example:
"You get started later in the day than #% of your classmates."

3/3

4: Average entry time

Example:
"You are active on average 2h 10m later in the day than your classmates."

Example:
"You are more studious / hard-working / social than #% of your classmates."

Sorting & comparing the data

An overview of the algorithms and their insights

Applying a custom sort order

Placing people into percentile brackets

Applying a custom sort allows me to view the rankings of each person according to any metric.

Calculating percentiles allows me to see the exact proportion of people that are above / below you in the rankings as a percentage.

It also allows me to find the median (middle) of the data, despite it being heavily skewed to one side.

Creating social nudges

Carefully observing the data allowed me to find insights

Insights can be turned into 'social nudges' by using comparative language:

Top 8% highest proportion of leisure

↓

"92% of your classmates recorded a lower proportion of leisure than you."

Creating a prototype

Using InVision Studio to create an interactive prototype