Chen Xinyu
Hello there! I am Xinyu, a researcher passionate about data science applications.
Research Interests
Urban Data Science, Machine Learning applications, Human-computer Interaction in accessibility
Education
-
M.S. Applied Geographic Information Systems National University of Singapore (August 2021) -
B.S. Human Geography and Urban-rural Planning Sun Yat-sen University (June 2019)
Work experience
Research Associate @ Urban Ecology Lab, National University of Singapore(February 2024 - present)
- Research on a novel method for measuring urban park accessibility: examining the ‘quarter-mile rule’.
Research Engineer @ School of Computing and Information Systems, Singapore Management University (December 2021 - December 2023)
Graduate Researcher @ Urban Analytics Lab, National University of Singapore (January 2021 - August 2021)
Teaching Assistant @ Urban Analytics Lab, National University of Singapore (August 2020 - November 2020)
- Prepared lecture materials about spatial network analysis by using QGIS
Research Projects
Smart Barrier-Free Access (SmartBFA)
HCI in urban accessibility
- Participated in developing a navigation tool to help wheelchair users find barrier-free accessible paths in Singapore
- Modified GraphHopper API by using JavaScipt to conduct map matching on GPS trajectories of accessible paths collected by the wheelchair volunteers
- Developed machine learning algorithms to detect the transition points between paths and open spaces for further routing generation by applying hidden Markov model and logistic regression
Improving Quantal Cognitive Hierarchy Model Through Iterative Population Learning (QCHIPL)
Game theory, Human Behavior Prediction
- Perform cross-validation on the developed Quantal Cognitive Hierarchy - Iterative Population Learning (QCH-IPL) model and compare the prediction results with baseline models
- Developed codes in Python to organize the entire workflow and conducted implementations of data preparation, model training, results comparison, and visualizations
Mining Real Estate Ads and Property Transactions for Building and Amenity Data Acquisition
Urban Science, GIS
- Developed an entirely new mechanism to maintain spatial databases by detecting unmapped buildings and amenities and adding unfilled building attributes detected from real estate data
- Extracted locations and attributes of buildings and amenities from 295,880 real estate listings and 954,510 transaction records by using Python and R
- The mechanism can achieve accuracy rates of over 90% in identifying locations of unmapped amenities and adding unfilled building attributes
Data engineering work
Gojek taxi data processing
Language: Python
Developed codes to aggregate over a million records of order-level taxi data into trip-level taxi data. For more details, please View codes.
Other coursework projects
Spatial-temporal Analysis of Taxi Demand in New York City
Urban Science, spatio-temporal analytics, transportation
View project report, codes, and data
Exploring Influencing Factors of Airbnb Penetration in Three U.S. Cities Using Spatial Regression Analysis
Urban Science, spatial analytics, real estate
Publications
- Yuhong Xu, Shih-Fen Cheng, and Xinyu Chen. Improving Quantal Cognitive Hierarchy Model Through Iterative Population Learning (Extended Abstract), 22nd International Conference on Autonomous Agents and Multiagent Systems (AAMAS-2023), London, UK, May 2023.
- Chen, X., & Biljecki, F. (2022). Mining real estate ads and property transactions for building and amenity data acquisition. Urban Informatics, 1(1), 12. [Urban Informatics Paper of the Year Award 2023 (1st place)]