MUCHEN ZHANG

Passionate data-driven researcher with a background in Sustainable Architecture and a current focus on Energy Systems and Data Analytics. Proficient in Python, GIS, and Machine Learning, with a proven record of applying advanced predictive modeling and multi-criteria spatial analysis to solve urban infrastructure and renewable energy challenges.

Next-Hour EV Charging Congestion Prediction (Machine Learning)

• Evaluated the performance of 4 models (Linear Regression, Random Forest, XGBoost, HGB) to predict station occupancy across 275 Traffic Analysis Zones (TAZs) in Shenzhen, China.
• Achieved a HGB model with PR-AUC of 0.95 and 91.1% Recall for busy-state detection, winning baseline models by effectively handling class imbalance.
• Engineered features for meteorology, POI density, and spatial connectivity while preventing temporal leakage via rolling-origin validation , to optimize grid load management strategies.

Multi-Criteria Spatial Analysis for UK Onshore Wind Farm Siting & Financial Analysis

• Conducted a GIS-based Multi-Criteria Decision Analysis (MCDA) to identify 10GW of optimal onshore wind farm sites across the UK.
• Modelled weighted site suitability map by integrating Wind Speed, Proximity to Grid/Road, Slope, Index of Multiple Deprivation (IMD), and Population Density. Proposed the selection of top 20 sites.
• Performed financial modelling (NPV, LCOE, CAPEX) to ensure economic viability.

Spatial Optimization of EV Infrastructure for Air Quality

• Evaluated the spatial efficiency of London's EV charging network by cross-analyzing NOx transport emissions data (LAEI 2022) against current charger distribution to identify most effective deployment zones.
• Processed and integrated multi-source datasets , including geocoding raw point data from the National Charge Point Registry (NCR) and performing spatial joining and coordinate transformations using Python and QGIS .
• Developed an interactive visualization pipeline using Folium , implementing MarkerCluster and heatmaps to reveal critical supply gaps in high-pollution corridors (e.g., West London highway/airport zones).
• Identified a strategic spatial mismatch , proving that current infrastructure is primarily market-driven rather than strategically focused on mitigating the highest pollution sources.

• Lead Department Rep at UCL BSSER Department
• Academic Rep of MSc Energy Systems and Data Analytics
• Academic Rep at University of Liverpool
• Secretary of Liverpool Debate Society