DataViz Emotion Capture – Eric Marshall

Background

In the modern world, data is everything. From farming to space exploration, there are few industries that do not rely on large amounts of data, and furthermore the representation of large data visually. One of the most common places for consumers to find data visualizations, however, is in news articles. As data becomes an ever more important facet of life, more and more news organizations are delving into the realm of data visualization. At this point, with so many things being represented, however, there has been very little study done into the reactions people have emotionally to the data visualizations.

Using new libraries that involve webcams and other sensors, however, may allow for programs to read (and possibly adapt to) the reactions that users have to the data they are viewing. By incorporating such technology into web pages that contain visualizations, companies would be able to further understand how the data being portrayed is affecting their user base. There is currently no method for gathering and utilizing such data, however, leaving the space open to innovation.

Executive Summary

By far the most common place for consumers to run across data visualizations is on the web. As such, the most extensive platform to pull data from would be within a web-based library that could utilize the computer’s webcam to analyze the emotions of the user as they are viewing a webpage-based data visualization. Because the goal of the system is to be able to actively gauge whether the user is looking at a visualization and then examine their emotional reaction, it will have to know when the user is actually looking at the visualization and be able to recognize when the user’s gaze has shifted, so that it can stop collecting emotional data. This will keep the collection to only data that actually pertains to the desired state, and means that there does not have to be someone who will pull only relevant data after the fact.

Once the data is collected, there will be a suite of visualizations and analysis that can then be applied to that data, which will help news organizations to further understand the impact of their data representations and improve them both in the future and in real-time, allowing for more effective representations and displays. By giving such a tool to news organizations and companies, we will be enabling them to create more meaningful graphics and information in the future.

Viability Analysis

There are lots of issues that we will come across, the first of which will be recognizing the position that a user is examining on the screen and recognizing when it is or is not a data visualization. Another issue that will be run into will be the actual recognition of emotions through the sensors that are available, and to associate that emotion with the currently viewed data visualization. Finally, we will need to figure out how the data can be best put into other visualizations so that it can be utilized by news organizations.

Risks and Rewards

The first and foremost of the issues associated with this project are primarily involving the privacy of the subjects. So long as there is a camera fixed on them gathering data, there is the possibility of sacrificing the user’s privacy. As such, all data involved in the system must be encrypted. Another, slightly more worrying risk of this project is the potential for its use in emotional manipulation. Because news organizations would be privy to the emotional states of their users as they view given data visualizations, it would then be more possible for them to figure out how to manipulate their viewers, opening the door to ethical dilemmas.

The rewards that could come from this, however, could revolutionize the news and data visualization industries significantly. By being able to adapt how data is represented based on users’ emotional reactions, they could learn to represent data in ways that will more meaningfully convey the information it represents.

Closing

In designing a web library to analyze emotion we could potentially unlock a whole new level of personalized informational media – data visualizations could potentially be manipulated to normalize people’s emotions and allow for more neutral representations – paving the way for a world of better, more personal information

Print PDF