Projects

You will be able to choose one of three types of projects, these will have different requirements regarding previous skills, and time commitment. In the past we have had teams with "free riders" who were not carrying their weight, leading to dissatisfaction and demotivation of contributing team members. We therefore encourage you to be upfront, and tell your colleagues how much work you want to invest in your project. The more you do, the more you will learn, but this also takes more time beyond the official requirements. Skills required for the different projects are:

  • Software development

  • Statistics

  • Web design/Information visualization

  • Community building/Marketing/Product design


COOLHUNTING PROJECTS

REQUIREMENTS:

  • Software development (medium)

  • Statistics (medium)

  • Web design (medium)

  • Community building (low)

  • Level of difficulty (low to intermediate)

Tracking the Attitude towards Vaccination and other unsubstantiated claims

Wikipedia provides an excellent mirror of social perception. The goal is to show the evolution of understanding of past events based on events in the present. For instance, the Anti-Vaxxers base their belief that vaccination causes autism on one single paper, which was later show to be statistically insignificant. Today, in the time of Covid19, there is frenetic discussion about conspiracy theories around vaccinations, track the origins and diffusion of conspiracy theories on online social media in the wake of Covid19.

Measuring the Ethics and Emotions of Tribes

Members of digital virtual tribes show their membership through their emotions in reaction to external events. For instance, depending on their ethical values, people will either react with compassion, or with anger and rejection to requests for support from refuges. Using the galaxyscope tribefinder system, build emotion tribes that will show their ethical and moral values (caring+fairness, or tradition+security) through the way they use words.

Predicting the Winners of Covid19

The day Moderna published the first results of its pre-clinical trials, it's stock price dramatically increased. It subsequently emerged that venture capitalists on the board of Moderna had sold substantial amounts of stock, reaping huge profits. Whether it is face masks, hand sanitizer, or vaccines, opportunities abound.

In an earlier similar instance, pharmaceutical company Valeant bought producers of successful drugs that covered niche diseases, boosting the drug prices enormously and garnering huge profits. Correctly predicting these scams led to huge profits of investors shorting the stock of Valeant.

Identifying predictors of negative Entanglement

The novel metric of entanglement measure the amount of collective awareness of an organization. Usually this is taken in the positive sense, as an entangled organization operates far more efficiently and productively. However, there are also negative examples, for instance the poisonous cultures of Enron or Lehman Brothers before they want bankrupt based on a culture of corruption and deceit. In this project, analyzing e-mail archives of Enron and other organizations, find metrics of negative entanglement

Predicting popularity of YouTube videos based on Facial Emotions

The success of YouTube videos depends (among various other factors) significantly on the choice of thumbnail and title. Therefore, Youtubers gain also financial profits from an attractive choice of these factors. To generate the dataset we start with a "pre-processing" step. In this step, criteria such as "ratio of views to average views of the channel videos", "ratio of views to channel subscribers", "average watch time of the video", etc. can be used to calculate an "attentiveness score". This would give us a dataset containing the thumbnail, title of the video (and category of the video) and the attentiveness score. In addition we can compute emotions of faces in the thumbnails and correlate them with the success.

PREDICTION/MACHINE LEARNING PROJECTS

REQUIREMENTS:

  • Software development (high),

  • Statistics (high),

  • Web design (low),

  • Community building (low),

  • Level of difficulty (high)


Predicting Personality from Watching YouTube Videos

Build a WebApp using facial emotion recognition. The idea is that when people watch YouTube videos, the emotions in their face in response to the video will reflect their personality and social and ethical values. To build the App, create a training dataset where people (the participants in this seminar) watch YouTube videos with emotional and/or divisive issues, and also enter their personality, risk attitudes, and ethical and moral values. Combine this with asking the same people to take four surveys online on the happimeter Web site.

Building an AI Matchmaking System based on Honest Signals

Using the methodology of the course, create a matchmaking system that extracts features from text and other honest signals, to identify significant others from e.g. Signal/Telegram traffic or other messages. The first challenge will be to find an suitable training dataset, e.g. Tweets of known (married) couples.

Measuring the Happiness of Plants

Using latest results from the International Laboratory in Plant Neurobiology, develop a plant sensing system using electric sensors which is capable of tracking the perception of plants and their communication with their environment measuring their signaling in their chemical and physical communication gateways. We are working with the Fintan foundation who uses biodynamic principles to communicate with plants, and measure the action potential of Mimosa Pudica and Codariocalyx Motorius, as well as of vegetables like beans to track human-plant interaction using the spiker box.

Measuring Emotions of Dogs and Horses through their Body Language

Extend the facial emotion reading system for dogs, developed in the COINs seminar of spring 2019, build a system on Android that labels the emotions of dogs or horses that you look at, based on their body language. You will have to collect libraries of labeled pictures of dogs or horses, where experts will have to agree on the emotion of the animal. For instance a wagging tail of a dog is correlated with joy.

We are working with horse researchers at University of Oklahoma that have decades of experience studying horse emotions.

Measuring Emotions in Zoom Meetings

Build an emotion tracking system based on facial emotion recognition and give feedback to meeting participants about how they feel during the zoom meeting. Additionally there is also the smartphone "meeting balancer" who measures who is talking (too much). The goal is to create a happier and more engaged meeting environment.

Tips for Groupwork

FINAL PAPER

The final paper should be 10 to 20 pages, plus references, coolhunting results, and code snippets when appropriate. The total number of pages also depends on the type of project, for a coolhunting project, the paper is typically longer, for a system development project, it can be shorter and basically just describe the system, for a statistics project it should document the main insights. Paper writing tips are in the presentation here.


Final Paper Structure:

  1. introduction - what is the problem

  2. related work - what have others done so far

  3. hypotheses - what are we trying to prove or disprove

  4. method - how are we doing it

  5. results - what have we found

  6. discussion - how does it compare to what others have already found

  7. weaknesses of the research, and possible future work