LTAT.06.010 Pervasive Data Science Seminar
Pervasive Data Science is a research field in the intersection of pervasive computing and data science that builds on the increased proliferation of sensors into everyday environments. These sensors produce large volumes of complex, real-time data streams, which are opening scientific investigations at an unprecedented scale. Examples of the adoption of pervasive computing and sensors abound, with air and aquatic pollution monitoring, patient health monitoring and autonomous drone systems, being examples of domains benefiting from pervasive data.
- Lectures: Narva mnt 18 - 2034
- Schedule: Monday 16:15 - 18:00
- Instructor: Huber Flores
The pervasive data science seminar provides an overview of techniques that can be used to analyze large scale sensor datasets, and principles to execute sophisticated machine and deep learning techniques in resource constrained devices, where processing power, energy and storage are limited. Besides this, the seminar also introduces to students with tools and methodologies for data sampling and acquisition, through users studies, surveys and other human-computer sampling methods. In addition, the seminar introduces beginner students to the design and development of experimental testbeds as well as guidelines to communicate their results to others successfully. Participants are expected to give oral presentations on a particular topic and solve practical tasks using R, MatLab or Python.
Announcements
- Lectures are in person in the lecture room (Contact the lecturer in case you missed a lecture or fall behind the course schedule.)
- Join Message Board
Examples of project demonstrations (Spring/Fall 2024)
Demo: Road quality estimation using accelerometer analysis: Watch here - Author: Joosep Näks
Demo: Systematic performance evaluation of counting people with low-cost sensors: Watch here - Author: Reo Kuchida
Demo: Wireless sensing for liquid identification: Watch here - Authors: Anette Habanen and Karl Raud
Poster: Transforming times series into images for deep learning: Author: Gregor Rehand
