physics.soc-phFeb 17, 2026

The Skeletal Trap: Mapping Spatial Inequality and Ghost Stops in Ankara's Transit Network

Authors: Elifnaz Kancan

Organizations: Independent Researcher, Ankara, Turkey

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Ankara's public transport crisis is commonly framed as a shortage of buses or operational inefficiency. This study argues that the problem is fundamentally morphological and structural. The city's leapfrog urban expansion has produced fragmented peripheral clusters disconnected from a rigid, center-oriented bus network. As a result, demand remains intensely concentrated along the Kizilay-Ulus axis and western corridors, while peripheral districts experience either chronic under-service or enforced transfer dependency. The deficiency is therefore not merely quantitative but rooted in the misalignment between urban macroform and network architecture. The empirical analysis draws on a 173-day operational dataset derived from route-level passenger and trip reports published by EGO under the former "Transparent Ankara" initiative. To overcome the absence of stop-level geospatial data, a Connectivity-Based Weighted Distribution Model reallocates passenger volumes to 1 km x 1 km grid cells using network centrality. The findings reveal persistent center-periphery asymmetries, structural bottlenecks, and spatially embedded accessibility inequalities.

Figures & tables

Figure 1: Urban Activity Proxy (Nighttime Lights): Heatmap illustrating Ankara’s leapfrog expansion and peripheral clusters outpacing traditional transit infrastructure in 2025. [ 10 ] .
Figure 2: Radar chart illustrating the distinct algorithmic fingerprints of the three grid categories, demonstrating that Critical Grids (red) uniquely exhibit simultaneous failure across both static anomaly detection (left) and dynamic regime stability (right) axes, while Anomaly Only and Regime Shift Only categories display mutually exclusive failure patterns
Figure 3: Daily Passenger Count Trends and Temporal Anomalies
Figure 4: Heatmap of daily occupancy rates for the 20 most volatile routes , highlighting extreme fluctuations
Figure 5: Spatial Distribution of Average Occupancy Rates: A grid-based visualization identifying extreme localized saturation (red zones) in peripheral districts such as Sincan and Batıkent, highlighting where demand exceeds available supply.
Figure 6: Grid-based Capacity Utilization Analysis: Mapping the percentage of used seat capacity across the network .
Figure 7: Spatial density of passenger demand across Ankara’s bus network , visualized as the average daily ridership per stop within 1 km² grid cells.
Figure 8: Spatial distribution of "Ghost" stops , identifying that 67.1% of the network consists of physically present but operationally inactive nodes.
Figure 9: Spatial distribution of anomalies and regime shifts identified by the six unsupervised algorithms: (A) Autoencoder anomaly detection (red) based on reconstruction error; (B) Isolation Forest (blue) identifying statistical outliers; (C) Local Outlier Factor (orange) highlighting density-based contextual deviations; (D) Graph-based spatial anomaly (purple) showing divergence from geographic neighbors; (E) Embedding Stability Analysis (teal) visualizing high daily geometric drift in latent space; (F) Daily Clustering Regime Shift (green) indicating high frequency of categorical switching between operational archetypes.
Figure 10: Regime shift map visualizing temporal instability: purple grids exhibit both geometric drift and frequent categorical switching, while blue and teal grids represent instability in only embedding space or cluster assignment, respectively.
Figure 11: Anomaly map highlighting service-demand misalignment: red indicates critical anomalies confirmed by three or more detection methods, orange shows moderate anomalies confirmed by two methods, and green represents normal operating conditions.
Figure 12: Spatial intersection of structural service-demand anomalies and temporal regime instability, highlighting 13 critical grid cells (red) where Ankara’s centralized transit planning fundamentally fails to accommodate the spatiotemporal complexity of urban sprawl.
AlgorithmDiagnostic QuestionType of Failure Detected
Autoencoder (PCA)Does this grid’s ridership–frequency–occupancy relationship conform to the system’s normative logic?Systemic logic failure : Locations whose supply-demand configuration cannot be reconstructed from the dominant latent patterns, indicating a breakdown in the system’s internal grammar.
Isolation ForestIs this grid statistically extreme in the multidimensional feature space?Extreme outliers : Grids occupying the far edges of the distribution, such as exceptionally high passenger loads paired with minimal trip frequency.
Local Outlier FactorDoes this grid deviate from its operational peer group?Contextual neglect : Locations that appear normal in absolute terms but are anomalously underserved relative to grids with comparable density profiles.
Graph-based SpatialDoes this grid’s service profile diverge from its geographic neighbors?Spatial discontinuity : Adjacent grid cells receiving dramatically different levels of service despite occupying the same urban fabric.
Table 1: Anomaly detection algorithms and their diagnostic perspectives on service-demand misalignment.
AlgorithmDiagnostic QuestionType of Failure Detected
Embedding StabilityDoes this grid’s operational identity (ridership–frequency profile) remain consistent across days, or does it drift erratically?Representational unpredictability : Locations that behave like a busy urban center on one day and a quiet suburban periphery on the next, preventing the system from establishing a stable reading of the grid’s identity.
Daily ClusteringCan the system consistently assign this grid to the same operational archetype each day?Categorical instability : Locations whose assigned cluster oscillates between high-demand, low-service (deficit) and low-demand, low-service (normal) from one day to the next, indicating that the system cannot commit to a stable classification.
Table 2: Regime shift algorithms and their diagnostic perspectives on temporal instability.
More in this topic

Urban Planning

Urban planning uses geospatial data, machine learning, and optimization to analyze cities and guide transport, land use, housing, and climate decisions.

See all 48 papers

Explore similar work

CardsList

  1. Public transit gains and spatially uneven travel demand changes after NYC congestion pricing

    Jun 16, 2026Donghang Li, Dingyi Zhuang, Yunlin Li +5Counterfactual EvaluationCausal Effect Estimation

    Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA · Department of Urban Studies and Planning, Massachusetts Institute of Technology, Cambridge, MA, USA · Center for Computational Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA +4

  2. Unequal Trips, Unequal Places: Diagnosing and Mitigating Delay Inequity in Autonomous Vehicle Fleet Coordination

    Jul 27, 2026Nicole Hu, Mingtao Zhang, Haoyang LI +2Algorithmic FairnessVehicle Routing

    School of Hotel and Tourism Management The Hong Kong Polytechnic University Hong Kong SAR, China · Department of Computing The Hong Kong Polytechnic University Hong Kong SAR, China · School of Hotel and Tourism Management & Computing The Hong Kong Polytechnic University Hong Kong SAR, China

  3. Transferable Human Mobility Network Reconstruction with neuroGravity

    Apr 26, 2026Jinming Yang, Shaoyu Huang, Zongyuan Huang +4Geospatial Representation LearningCross-Domain Transfer Learning

    MoE Key Laboratory of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University, Shanghai 200240, China · Data-Driven Management Decision Making Lab, Shanghai Jiao Tong University, Shanghai 200240, China · Department of City and Regional Planning, University of California, Berkeley, CA 94720, USA +2