Skyhook is excited to be partnering the Singapore University of Technology and Design (SUTD) to carry out Singapore’s first large-scale National Science Experiment. This experiment is organized by the Ministry of Education and the National Research Foundation Singapore to develop an interest in science among the country’s young people. In this project, fifty thousand students across Singapore will carry a sensor to collect data on travel mobility patterns, location, environment, etc. The students learn how to analyze the data as part of its Science, Technology, Engineering and Mathematics (STEM) program.
We spoke with Nils Ole Tippenhauer and Erik Wilhelm, both Assistant Professors at SUTD for more information about this research project:
1. Tell us about your work at Singapore University of Technology and Design - what is this project you are working on and what are you looking to achieve?
I am an Asst. Prof. in the Pillar of Information Systems Technology and Design at Singapore University of Technology and Design (SUTD). Together with my colleague Erik Wilhelm, I have been involved in the "National Science Experiment" project since September 2014. The National Science Experiment is a nationwide initiative to spur interest in STEM subjects in Singapore students. To achieve that goal, we designed and built 50,000 personal sensors (which we call SENSg, pronounce SENSE-SG), together with our project partners, such as Delta Electronics.
The SENSg devices are worn by Singapore students over a week to measure their daily routine. For example, the devices track the number of steps the students take each day, the amount of time they spend indoors and outdoors, and their travel patterns. Each student’s data is automatically aggregated in our cloud-based infrastructure, and accessible by the student through a web application. In addition, the devices record environmental parameters such as temperature, pressure, humidity, noise level, light level, infrared temperature, and motion. Students can view each measurement, compare aggregated data among their cohort, play a game to encourage more daily activities, and create statistical challenges using their personal data. Their data can be downloaded to enable teachers to explore the concept of big data and its applications in the real world with their students in the classroom.
The combined data of participating students is analyzed to provide insights into the travel mobility of Singapore youths for decision-making at the urban system-level (e.g. bus schedule optimization).
2. What are you most excited about in this project? What do you hope to learn about?
This unique opportunity to collaborate on a sensing project on a nationwide scale, together with its knowledge learning opportunities, is exciting for us all. Our main motivation is to present the new world of connected devices and people in a tangible manner to Singapore students. After the experiment, we hope gain big data insights from these findings.
3. How are you leveraging Skyhook’s location data in your project?
One major challenge of this project is to keep the individual sensor node price down and fit all of its sensors into a compact form factor. At the same time, we needed to tag each set of measurements in space and time. From the start, we decided that GPS would be too expensive in terms of hardware and battery life. In addition, we expected the students to be indoors a fair amount of time during the day, so we require a solution to provide indoor localization as well. Skyhook’s localization system was sufficient for our needs. The sensor nodes record location "fingerprints" of neighboring WiFi networks with each measurement. Those fingerprints are then later used to find the location of the measurement.
Our initial prototypes showed good accuracy of the localization results returned by Skyhook. Additionally, Skyhook has also been supportive in technical matters. We were initially concerned about the massive amount of queries we would have to perform (5,000 per second) to ensure nearly real time data availability for the students, but we managed to work with Skyhook to implement solutions for this problem.
4. How do the students go about analyzing the data being collected? Any unexpected challenges so far? What is the outcome of their project, or what do you hope that they will learn?
The students can analyze their data with our interactive web portal, or download the raw data. On the web portal, the data is visualized in charts and a map that leverages the measurement location. In addition, the data can be explored interactively through several games and interactive personal data activities.
Overall, we hope to expose students to physical measurements such as air pressure, humidity and ambient noise. In addition, we hope to encourage the students to explore and analyze their collected data for insights about their daily life, and their environmental impact such as CO2 footprint.
5. Our understanding is that the project launched in August, is currently taking place and it has a timeframe of 3 years. In terms of preliminary data collected, what have you seen that has been interesting so far? Any notable trends or unexpected information?
Using the localization technology, we were able to trace the participant’s travel paths, and even find locations of missing devices. One surprising finding is that some of the students commute from outside the country every day!
The overall experiment is conducted in several stages, of which three main ones started in September. In our preliminary analysis of the data, we observe that students’ activity spans a long period of the day (some students are up to 16 hours on the move), and that during their examination period, they are typically indoors and studying.
6. You are collecting data on location, weather, and speed. Any other data you are collecting? Are you going to tie all this information together? If so, how?
In total, we are collecting information on temperature, ambient audio noise, ambient light, humidity, air pressure, WiFi networks, steps, and current travel modes. We combine all of them in the general analytics framework that we set up for this experiment. In addition, some of that information has proved useful for us to manage logistics of the sheer amount of devices and schools involved in this project.
7. What are the roadblocks that you foresee “big data” having to overcome to be an integrated and more popular part of how we process information?
Big data is always a challenge for tools and infrastructure. Solutions for both are relatively novel, and need to be learned by users. We hope that this project will motivate the participants to learn more about these topics, and become the big data experts of tomorrow!
8. Anything else we should know to understand the motivations behind your project, or links to further information for interested readers to learn more?
General information can be found on the experiment at the website, www.nse.sg.
About Nils Ole Tippenhauer:
Nils Ole Tippenhauer is an Assistant Professor at the Information Systems Technology and Design Pillar at the Singapore University of Technology and Design. He earned his Dr. Sc. In Computer Science from ETH Zurich (Switzerland) in 2012. Before going to ETH, Dr. Tippenhauer received a degree in Computer Engineering from the Hamburg University of Technology (Germany) in 2007. He is interested in IoT infrastructure systems, wireless localization, and information security aspects of applied systems.
About Erik Wilhelm:
Erik Wilhelm is an Assistant Professor in the Engineering Product Development Pillar at the Singapore University of Technology and Design. He earned his PhD from the ETH-Zurich while studying multi-criteria vehicle design, data analytics, and control optimization. While in Zürich, Dr. Wilhelm co-founded a start-up in the area of vehicle telematics for reducing onroad energy use. His post-doctoral research was performed at the Massachusetts Institute of Technology in the Field Intelligence Lab. Erik Wilhelm’s research goal is the advancement of economically and ecologically sustainable transport.