Crowdsourced reviews reveal substantial disparities in public perceptions of parking

TL;DR

The study tackles the challenge of measuring public perception of parking by leveraging crowdsourced Google Maps reviews at a national scale. It combines BERT-based sentiment classification with Lexical salience-valence analysis and employs Generalized Additive Models to relate parking sentiment to a rich set of socio-spatial covariates at CBSA and CBG levels, revealing substantial urban-rural and demographic disparities. Key findings show that sentiment strongly depends on POI type and urban context, with denser, diverse, and economically stressed areas tending toward more negative perceptions, and simply increasing parking supply not guaranteeing improved sentiment. The work demonstrates the utility of crowdsourced textual data for hyperlocal parking diagnostics and offers policy directions such as targeted demand-management, equitable pricing, and improved accessibility, while acknowledging biases and data limitations. These insights can guide targeted parking management strategies and future research using complementary data sources and longitudinal analyses.

Abstract

Due to increased reliance on private vehicles and growing travel demand, parking remains a longstanding urban challenge globally. Quantifying parking perceptions is paramount as it enables decision-makers to identify problematic areas and make informed decisions on parking management. This study introduces a cost-effective and widely accessible data source, crowdsourced online reviews, to investigate public perceptions of parking across the U.S. Specifically, we examine 4,987,483 parking-related reviews for 1,129,460 points of interest (POIs) across 911 core-based statistical areas (CBSAs) sourced from Google Maps. We employ the Bidirectional Encoder Representations from Transformers (BERT) model to classify the parking sentiment and conduct regression analyses to explore its relationships with socio-spatial factors. Findings reveal significant variations in parking sentiment across POI types and CBSAs, with Restaurant POIs showing the most negative. Regression results further indicate that denser urban areas with higher proportions of African Americans and Hispanics and lower socioeconomic status are more likely to exhibit negative parking sentiment. Interestingly, an opposite relationship between parking supply and sentiment is observed, indicating increasing supply does not necessarily improve parking experiences. Finally, our textual analysis identifies keywords associated with positive or negative sentiments and highlights disparities between urban and rural areas. Overall, this study demonstrates the potential of a novel data source and methodological framework in measuring parking sentiment, offering valuable insights that help identify hyperlocal parking issues and guide targeted parking management strategies.

Figures (9)

• Figure 1: Descriptive results for parking sentiment by POI. (a) Number of POIs by types in the collected dataset (top 10). (b) Distribution of POI count and number of reviews. The x-axis represents the log(reviews), while the y-axis represents the POI count. A threshold of 10 is applied to plot the distribution. (c) Distribution of sentiment. The x-axis represents the weighted sentiment, calculated as the average of all sentiments associated with a POI, while the y-axis represents the POI count. POIs with fewer than 10 parking-related reviews are excluded from the analysis. (d) Pairwise Wilcoxon testing between the sentiment distributions of two POI types.
• Figure 2: Distribution of parking sentiment across CBSAs in the U.S. (a) Nationwide distribution of parking sentiment, with CBSA boundaries outlined in black. Areas with negative sentiment are shown in blue, while areas with positive sentiment are shown in red. (b) Boxplot of average parking sentiment for the top and bottom 10 CBSAs, including only those with more than 50 CBGs. (c) Close-up views of the two CBSAs with the most positive and negative parking sentiment, using the same color scheme as in panel (a).
• Figure 3: Correlation between parking sentiment and socio-spatial factors, with only correlations having a P-value $<$ 0.001 displayed. (a) Within-CBSA correlation. (b) Between-CBSA correlation segmented by POI types. (c) Scatter plot of univariable regression showcasing the four strongest correlations by POI types, where each point represents a CBSA, sized according to its total population.
• Figure 4: Textual analysis using LSVA based on (a) all POIs, (b) POIs located in the urban areas, and (c) POIs located in the rural areas. Terms on the extremes of the x-axis indicate high or low salience in the reviews, while terms on the extremes of the y-axis indicate strong positive or negative sentiment.
• Figure 5: Classification performance of candidate models: Precision, Recall, F1-score, and Training and Testing accuracy.