Abstract This study analyses 16,969,327 Metropolitan Police street-level crime report records covering December 2010 to July 2026, obtained from the UK Police Open Data platform. After cleaning (removing records with missing spatial identifiers and duplicates), a final analytical dataset of 14,245,882 records across 16,110 Lower Super Output Areas (LSOAs) and 16 crime categories was produced. The research integrates exploratory data analysis, LSOA-level spatial aggregation, time-series visualisation, and interpretable machine learning using a Decision Tree Regressor. A temporal train-test split was employed, using records before July 2024 for training and the final twenty-four months for testing. This split was chosen, as opposed to a random one to avoid data leakage and as a factor of 12 months to address the seasonality of crime. Model performance was compared against a naive baseline (mean crime count per LSOA-crime type combination, computed from the training period). The Decision Tree achieved a Mean Absolute Error of 0.242 crimes per LSOA-month, outperforming the naive baseline’s MAE of 0.266, indicating that for long-term crime pattern predictions, tree-based regressors can provide a better performing tool for police resource allocation than simple averages. Feature importance analysis revealed that LSOA location plays the most important role, followed by temporal features. The crime type features were observed to have the lowest importance. The framework emphasises reproducibility, transparency, and operational applicability within a public-sector context.
With the rapid expansion of urban vehicle scale, traffic flow and congestion prediction have become increasingly important. To address the issue of inaccurate prediction of traffic flow and congestion, a graph convolutional network is proposed, and a gate controlled recurrent unit is introduced to design a spatio-temporal graph convolutional network. The network is employed to predict traffic flow in smart cities, and traffic flow prediction in smart urban environments model grounded on spatio-temporal graph convolutional network is constructed. At the same time, based on the flow prediction results, an improved fuzzy C-means clustering algorithm incorporating entropy weight method is introduced to establish a traffic congestion prediction model. The performance comparison analysis of the developed traffic flow prediction model with other models shows that its average mean square error and precision are 1.176 and 97.99
This study proposes a quantitative framework for assessing narrative environments in historic districts by integrating space syntax with agent-based simulation (ABS). Taking five historic districts in the Old Town of Xun County, China, as the case study, the research combines global integration, local integration, choice, and intelligibility to identify spatial accessibility, corridor potential, and cognitive efficiency, and uses ABS together with manual pedestrian observations to validate movement concentration and detect anomalous areas. Syntactic analysis identifies East Housi Street and South Street as the primary movement corridors, with the highest accessibility (global integration = 1.182 and 1.256, respectively) and pedestrian concentration, whereas Xinzhuang Street exhibits critically weak connectivity (intelligibility R2 = 0.427), while Pifu Road emerges as an important southern corridor (choice = 45,369) whose continuity is constrained by spatial obstructions. The results reveal a clear north–south gradient in spatial accessibility: northern districts exhibit higher accessibility and stronger pedestrian concentration, but also greater risks of over-commercialization and narrative homogenization, whereas southern districts are characterized by corridor fragmentation, weaker accessibility, and disrupted cultural continuity. ABS validation reveals two anomalies: North Xiaoximenli (low flow despite moderate syntax, blocked by illegal structures) and Nanshan (high traffic despite low syntax, anchored by the City God Temple). Based on these spatially explicit diagnostics, the proposed framework provides an evidence-based priority pathway for removing physical obstructions, reactivating southern heritage nodes, and restoring corridor continuity along East Housi Street, offering a replicable technical pathway for spatial diagnosis, narrative restoration, and conservation-oriented renewal in multi-district historic environments.
Educational systems must accommodate demographic concentration in cities while maintaining teaching capacity, yet whether student enrolment, teacher inputs, and school facilities respond uniformly to city size remains unclear. Here, we analyzed the scaling of these three dimensions with population across Chinese cities from 2000 to 2020. In 2020, all six indicators scaled sublinearly with population (β 0.74–0.87). In primary education, students and teachers shared an exponent of 0.87, whereas schools scaled more sublinearly (β = 0.74). These results indicate that average school size increases with city size, particularly in primary education, while student–teacher ratios remain broadly scale-neutral. Economies of scale in education provision therefore arise mainly through the organization of school facilities rather than through reduced teacher provision per student. Scale-adjusted indicators also revealed pronounced regional disparities: primary education indicators generally exceeded population-based expectations in central and southern China but fell below expectations in many northeastern, northern, and eastern coastal cities. The broadly corresponding spatial patterns of students, teachers, and schools suggest that educational resources generally follow regional differences in student enrolment. Overall, larger cities organize basic education through relatively fewer and larger schools without a corresponding reduction in teaching capacity relative to student enrolment.
