Los Angeles Parking Citation Analysis
Temporal, Spatial, and Predictive Patterns in Parking Enforcement, 2021–2025
2021–2025 · City of Los Angeles
Parking violations are structured, not random.
This web report provides an overview of patterns in Los Angeles parking citation data, focusing on time trends, location patterns, vehicle characteristics, and predictive results.
Exploratory Analysis
Spatial Clustering
Decision Tree
XGBoost
9.23M
raw citations
93.18%
retained after cleaning
$74.04
average fine
Overview
The analysis is based on cleaned Los Angeles parking citation data and examines major violation types, changes over time, spatial concentration, and underlying predictive patterns.
8.61M
complete observations
After listwise deletion and basic plausibility checks.
13
selected variables
Reduced from the raw dataset for analysis efficiency.
0.064
Cramer’s V: body type
Statistically significant, but practically weak.
0.070
Cramer’s V: vehicle make
Vehicle make has limited explanatory strength.
Data Cleaning
The raw records are reduced through variable selection, listwise deletion, and geographic plausibility checks.
Fine Amounts and Violation Types
Fine amounts are concentrated in a relatively narrow range between $60 and $80.
Interactive Figure 1. Vehicle Make and Body Style Composition.
This interactive figure compares the most common vehicle makes and body styles in the cleaned citation data. Hovering over each slice shows the exact share for each category.
This interactive figure compares the most common vehicle makes and body styles in the cleaned citation data. Hovering over each slice shows the exact share for each category.
Key point: citation activity is not evenly distributed across violation categories.
A few categories explain a large share of total citations.
Temporal Pattern
Monthly citation counts fluctuate across 2021–2025 rather than following a simple monotonic trend.
Interactive Figure 2. Monthly Parking Citation Counts, 2021–2025.
This interactive line chart shows how citation counts change over time. The range slider allows readers to zoom into specific periods and compare short-term seasonal movement.
This interactive line chart shows how citation counts change over time. The range slider allows readers to zoom into specific periods and compare short-term seasonal movement.
Monthly citation counts vary noticeably, with a clear peak in early 2022 followed by repeated increasing and decreasing pattern. It suggests cyclical or seasonal effects rather than a steady long term linear trend.
Spatial Concentration
Citations are geographically clustered, especially around dense urban and commercial areas.
Interactive Figure 3. Grid-Based Spatial Concentration of Parking Citations.
This interactive map shows where citations are geographically concentrated across Los Angeles. Larger and darker markers represent grid cells with more citation records.
This interactive map shows where citations are geographically concentrated across Los Angeles. Larger and darker markers represent grid cells with more citation records.
Interpretation
The spatial pattern suggests that parking violations are concentrated in dense urban regions, especially near downtown Los Angeles and other commercial zones.
Downtown LA
Commercial density
Parking demand
Enforcement intensity
Reading the map
Larger and darker markers indicate grid cells with more citations. Hover over each point to inspect the local citation count.
Predictive Modeling
Violation groups are predicted using temporal, spatial, vehicle, and plate-related features. XGBoost improves over the interpretable Decision Tree baseline.
Main modeling result
The pruned Decision Tree reaches 0.549 test accuracy and 0.457 macro-F1. XGBoost improves performance to 0.643 test accuracy and 0.570 macro-F1.
This indicates useful predictive signal, but some violation categories remain difficult to separate with the available variables.
Feature interpretation
The strongest predictors are related to time, plate status, location, and vehicle grouping. This suggests that violation type is shaped by a combination of enforcement timing, urban geography, and vehicle characteristics.
Conclusion
Parking violations in Los Angeles reflect structured urban patterns rather than random behavior.
Temporal
Citation counts vary over time, with recurring monthly movement and no simple linear trend.
Spatial
Violations cluster in dense urban regions, especially around central and commercial areas.
Modeling
XGBoost outperforms the Decision Tree baseline, but prediction remains moderate.
Limitation: the data reflect enforcement activity rather than all true parking violations.
Missing data removal and the absence of external variables may also limit interpretation.