Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
What can Smartphone Apps learn about their Users, using Big Data and Machine Learning?
Our vision is to create smartphone apps that know their users well and can therefore support them with their personal development goals or dealing with chronic illness.
Are you interested in exploring how well exactly your smartphone could get to know you?
Keywords: social computing, internet of things, android, big data, data analysis, machine learning
Smartphone apps can get access to data from various sensors as well as logs of user interaction, which allow inferences about users’ behavioral patterns.
For example, the number of sensed Bluetooth devices is an indicator of the time spent around other people. Similarly, frequent use of a calendar app indicates a high effort in planning.
Furthermore, we are interested in making inferences from behavior to the psychological states (e.g. emotions) and characteristics (e.g. personality traits) of the user. For example, from a large amount of time spent around other people we might infer a high level of sociability, and a high effort in planning is indicative of being well-organized.
Smartphone apps can get access to data from various sensors as well as logs of user interaction, which allow inferences about users’ behavioral patterns.
For example, the number of sensed Bluetooth devices is an indicator of the time spent around other people. Similarly, frequent use of a calendar app indicates a high effort in planning.
Furthermore, we are interested in making inferences from behavior to the psychological states (e.g. emotions) and characteristics (e.g. personality traits) of the user. For example, from a large amount of time spent around other people we might infer a high level of sociability, and a high effort in planning is indicative of being well-organized.
Your goal is to create a smartphone app that can sense some aspects of the user’s behavior.
The procedure of this Master thesis would be as follows:
1. briefly review the literature on smartphone sensing
2. decide on a set of data sources and associated behavioral variables that you find interesting
3. develop an Android app that collects the relevant data
4. deploy the app in a small field study, where participants provide ground truth information
5. develop and evaluate a predictive model that can predict the ground truth from the sensor data
Possible behavioral variables of interest are:
- time spent interacting with other people vs. being alone
- time spent at parties
- time spent at home vs. work vs. in public places
- time spent at a gym
- time spent with focused work
- time spent procrastinating
You should bring:
- Programming skills, preferably in Java (C/C++ and Objective-C are also valuable)
- Basic skills and interest in data analysis and machine learning
- Interest in psychology and human behavior
Your goal is to create a smartphone app that can sense some aspects of the user’s behavior.
The procedure of this Master thesis would be as follows:
1. briefly review the literature on smartphone sensing 2. decide on a set of data sources and associated behavioral variables that you find interesting 3. develop an Android app that collects the relevant data 4. deploy the app in a small field study, where participants provide ground truth information 5. develop and evaluate a predictive model that can predict the ground truth from the sensor data
Possible behavioral variables of interest are:
- time spent interacting with other people vs. being alone - time spent at parties - time spent at home vs. work vs. in public places - time spent at a gym - time spent with focused work - time spent procrastinating
You should bring:
- Programming skills, preferably in Java (C/C++ and Objective-C are also valuable) - Basic skills and interest in data analysis and machine learning - Interest in psychology and human behavior