Shuhao Li
Geospatial Data Science / Spatial Environmental Science / Machine Learning
Wenzhou, Zhejiang Province, China
Email: lishuhao010825@163.com
Msc @ University of Bristol
My Projects
For specific code implementation, please refer to the GitHub link: https://github.com/Coldzera-010825/geospatial-Lab
A Study of Urbanisation Trend Prediction in Nanchang Based on the CA-Markov Model
基于 CA-Markov 模型的南昌市城市化趋势预测
Used the CA-Markov model to simulate and forecast the land use of Nanchang city in 2030 and explore the urbanisation process of the city, to promote the healthy development of Nanchang city in the future and maintain the virtuous cycle of urban ecosystem
Analysis of soil erosion changes and influencing factors based on the CSLE model and GeoDetector in Dongjiang River Basin of China
基于 CSLE 模型和 GeoDetector 的土壤侵蚀分析
Quantified the spatial and temporal patterns of soil erosion in the area from 2016 to 2020 based on the Chinese Soil Loss Equation (CSLE), GIS and mathematical statistics. Employed multi-temporal Landsat imagery to derive vegetation cover and topography inputs, and used the GeoDetector method to assess the contribution of rainfall, slope, and vegetation indices to erosion variability.
Analysis of spatial and temporal variations and differences in PM2.5 intensity in urban and rural areas of Ganzhou City and study of its influencing factors
赣州市PM2.5时空尺度的城乡差异分析与因素研究
This study takes Ganzhou city as an example to compare the accuracy of three interpolation methods, and then analyze the characteristics and patterns of urban and rural PM2.5 concentration distribution from spatial and temporal scales. Finally, explores the spatial mechanisms of its various influencing factors using the geodetector method.
Bristol urban safety analysis and planning of students' routes to school
布里斯托城市安全性评估与学生安全上下学安全路线规划
This research focuses on enhancing urban safety and route planning for students in Bristol through a twofold approach. Firstly, it employs the U-Safety system, which integrates multi-source data, such as crime statistics, traffic data, and spatial features, with machine learning techniques(SAE, ANN and SVM) to calculate and predict a Safety Index (SI). Secondly, the study utilizes graph-based route planning algorithms, including Dijkstra's and A* algorithms, to design safer routes for university students.
Who is the Top Dog?
机器学习在犬种人气分析中的应用
Used Python (sklearn, statsmodels, seaborn, etc.) for regression and classification to identify factors influencing dog breed popularity and provide data support for selecting the best "internet-famous dogs." Applied machine learning techniques such as linear regression, logistic regression, random forest, KNN, decision tree, and cross-validation.
HCV Incidence Analysis among People Who Inject Drugs in Georgia (Team Research Project)
格鲁吉亚注射毒品人群中丙型肝炎发病率分析（团队研究项目）
Collaborated with public health students to analyze over 1 million records related to Hepatitis C Virus (HCV) testing data among injecting drug users. Cleaned and transformed the dataset using R, calculated incidence rates based on seroconversion (first negative, later positive), and conducted descriptive and survival analysis to examine population-level risk factors. Created geospatial visualizations to map incidence distributions and identify high-risk regions, contributing to epidemiological insights for public health planning.
Research on Spatiotemporal Characteristics of Carbon Emissions in the Pearl River Basin (Co-author, Land Degradation & Development, 2025)
珠江流域碳排放时空特征研究
Collaborated on constructing a high-resolution carbon emission grid by integrating land-use data with biomass burning records (2001–2020). Applied interpretable machine learning (SHAP) to identify natural and anthropogenic drivers, revealing forests as the main carbon sink (>99% sequestration) and rapid growth of emissions from construction land (+175.9%). Contributed to data processing, spatial analysis, and result interpretation, providing insights for carbon neutrality strategies and land-use planning.