Welcome to the course CS-C4100 - Digital Health and Human Behavior!¶
Course Description¶
The primary audience of this course is BSc students in their final year and MSc students with knowledge of Python and data science. You can find the course and enroll on Sisu by searching for the course code “CS-C4100”. This is a 5-credit course over period II.
Overview¶
In the past two decades, connected devices like smartphones, smartwatches, and online platforms like social media have become essential to our daily lives. Interacting with these devices and platforms, we leave digital traces behind. We often hear companies accessing these digital traces (our data) can learn much about our lives and behavior. But what exactly can be learned from these data, how it can be learned, and for what purposes can it be used? In this course, we learn about different kinds of digital traces that people leave behind. Working with data in some case studies, we learn how to extract from these data. We also discuss different ways to use the extracted information from the data. We focus on use cases that are related to health and well-being (e.g., self-tracking, monitoring people’s sentiments during crises on social media, and digital phenotyping).
Learning outcomes¶
After completing this course, you will learn how digital data from people (e.g., data from their smartwatches) can be used to measure different aspects of their behavior at individual and group levels. You will also learn how this information can be used in health and public health domains. By the end of the course, you will know what kind of questions can be answered using such data, and you will be able to implement code to extract relevant and valuable information from such data.
Description¶
This course will cover topics related to quantifying human behavior, health, and well-being using digital traces at the individual and group/society levels. The topics include 1) quantified-self, 2) digital data collection and analysis of behavioral and/or physiological data in the wild as well and 3) controlled digital data collection via recruitment of study participants for the purpose of quantifying behavior, health, and well-being, 4) digital phenotyping, 5) digital phenotyping for public health and 6) ethics, privacy and validity assessment in the context of all the topics covered in the course.
Prerequisites¶
Mandatory
Basic mathematics courses (especially introductory statistics)
Python programming skills
At least one introductory course in machine learning, data science, or artificial intelligence
Helpful to know
Python libraries such as Scikit-learn and Seaborn
Schedule¶
Wed 25.10.2023 14:15–16:00
Wed 01.11.2023 14:15–16:00
Wed 08.11.2023 14:15–16:00
Wed 15.11.2023 14:15–16:00
Wed 22.11.2023 14:15–16:00
Wed 29.11.2023 14:15–16:00
All lectures are held on campus (Location: R008/213a, Otakaari 4).
Exercise Sessions¶
Mon 30.10.2023 14:15–15:45
Mon 06.11.2023 14:15–15:45
Mon 13.11.2023 14:15–15:45
Mon 20.11.2023 14:15–15:45
Mon 27.11.2023 14:15–15:45
Mon 04.12.2023 14:15–15:45 (Questions related to the final project)
All exercise sessions are held online.
Workload structure¶
Weekly lectures (bonus attendance points)
Weekly exercises; data analysis
Weekly reading + multiple choice questions
Final project + Peer grading
No exam!
Lecture 1: Intro and background¶
Course overview and practicalities
Health, well-being, and digital data
Data sources
Types of data
Lecture 2: Quantified-self¶
What is quantified-self? How prevalent is it?
What is measured and how is it measured? (apps vs devices)
Applications of self-tracking
Lecture 3: Quantifying behavior of groups of people/sub-populations¶
Data collection in the wild/Social media data: case study: sentiment analysis during crises
Controlled data collection: case study: Reality mining/DTU/ identifying chronotypes
Lecture 4: Digital Phenotyping¶
What is digital phenotyping?
Areas where it can be used
Case-study: MoMo-Mood
Resources
Onnela, J. P., & Rauch, S. L. (2016). Harnessing smartphone-based digital phenotyping to enhance behavioral and mental health. Neuropsychopharmacology, 41(7), 1691-1696.
Lecture 5: Digital Phenotyping for Public Health¶
Resources
Waring, O. M., & Majumder, M. S. (2020). Introduction to Digital Phenotyping for Global Health. In Leveraging Data Science for Global Health (pp. 251-261). Springer, Cham.
Lecture 6: Privacy, ethics, assessment¶
Privacy and ethics in data collection
Data handling and safekeeping, GDPR
Assessment and evaluation of tools and interventions which are based on patient-generated data
Requirements & grading¶
Exercises are graded automatically/by the TAs
In-class tasks (not graded)
Projects are graded by peer review and teaching assistants
No exam
Assignments¶
Weekly assignments on Jupyer Hub to support lecture material