- The document discusses a study conducted by researchers at the University of Portsmouth analyzing anonymized Wi-Fi location data from students and comparing it to their academic performance. - The researchers aimed to explore student movements on campus and relate their locations over time to their grades, to make recommendations about where students can spend time to improve grades. - They processed Wi-Fi log data, timetable data, and academic data and found library time and lecture time were positively correlated with grades, while open access area time was negatively correlated.

1. Presented by
Soumabha Sarkar, Dr. Bryan Carpenter, Dr. Mohamed Bader and Arron Knight
University of Portsmouth, UK
Where Students Go and How They Do: Wi-Fi
Location Data Versus Academic Performance

2. • Initialization
• Wi-Fi Data
• Research Objectives
• Data Processing
• Application Output
• Future Work
Content

3. • University of Portsmouth is part of the EDUROAM community.
• University’s Information Services department have been collecting
wireless data on all Wi-Fi connections.
Initialization

4. Wi-Fi AMP Log Data
• Provided by Information Service Department, University of Portsmouth
• Total 31 different types of Information.

5. Wi-Fi AMP Log Data

6. What is Device Name
LION GATE
Building
Second
Floor Plan
LION-2-000
LION-2-001
.
.
.
LION-2-007

7. Data Anonymousness
 Ethical conundrum
Wi-Fi AMP log files contains Username information which is considered
as personally identifying information (PII). As well as the MAC address of
user's device is also sensitive information.
❑Solution
Developing an anonymization tool to hash username information and
MAC address information in data sheet. HMAC-SHA256 is used as the
hashing algorithm.

8. Useful Wi-Fi Data Fields

9. • Explore student movements within University Campus
• Relate students’ whereabouts over some period of time to how well
they do academically.
• Make evidence based recommendations to students about where best
to spend their time (library, lectures etc.) to improve their grades.
Research Objective

10. Time-Table Data
• File name pattern UNIT-Group-type-Day-StartTime-Room-register.csv
• Corresponds to a weekly teaching session that was running in the
period corresponding the Wi-Fi data.
• Contain a list of Student IDs.
Academic Data
• Contain a list of Student IDs.
• Marks obtained in UNITs.

11. Process Pipeline

12. Wi-Fi Data Processing

13. Application Output

14. Application Output

15. Application Output
• Library Time Positively Correlated
• Lecture Time Positively Correlated.
• Open Access Area time Negatively
Correlated.

16. • Consider Wi-Fi data for whole academic year.
• Need refining in terms of calculating time spent by students in a
particular location.
• Consider other statistical model for analysis and prediction.
• Consider the fact that not all the students have Wi-Fi enabled devices.
Future Works

17. Thank You