Truck idling and parking data for AB 617 disadvantaged communities study

This project investigates air pollution in California communities disproportionately affected by their proximity to transportation corridors, industrial facilities, and logistics centers, focusing on truck-related activities, including idling, parking search, and parking demand, using comprehensive datasets and robust models employing techniques such as Random Forest, Convolutional Neural Network, Bayesian Ridge Regression, and Spatial Error Model. Key findings reveal factors affecting idling times, parking search times, and parking demand, with heavy-duty trucks having the highest idle times and parking search challenges concentrated around transportation arteries and freight yards. The Spatial Error Model highlights relationships between truck activities, socio-economic variables, and air pollution in AB 617 communities. Based on these findings, preliminary policy recommendations include targeted anti-idling campaigns, improved truck parking facilities, cleaner fuels and technologies, enhanced routing efficiency, stricter emission standards, and strengthened land-use planning.

Methods
The data submitted in this dataset originates from various sources, with each source providing unique insights into the study of truck idling and parking in AB 617 Disadvantaged Communities. The dataset submitted here is the result of careful processing and manipulation of the original datasets to create a comprehensive view of truck idling and parking behaviors.

1. Geotab Ignition Platform Data
Though not directly included in this submission due to licensing restrictions, data from the Geotab Ignition platform was instrumental in the creation of this dataset. It includes raw idling data, raw data for searching for parking, and raw truck parking location data. We used these datasets to extract key metrics related to truck idling and parking behaviors.

The Geotab data was processed and aggregated to obtain daily idling times and parking search times. This was done by using the geohash provided to group data by location and then computing the daily metrics. Please note that due to licensing restrictions, the raw Geotab data is not included in this submission. For those interested in using the Geotab data, please refer to the Geotab website to access the data directly.

2. CalEnviroScreen 4.0, Census data, and OpenStreetMap (OSM)
These datasets provided contextual information, such as demographics and infrastructure, which were used to enrich the idling and parking data derived from the Geotab datasets. For example, demographic data from the Census and CalEnviroScreen 4.0 was used to identify disadvantaged communities, while data from OpenStreetMap was used to map idling and parking behavior to specific locations.

3. Kern County Traffic Count Data System (TCDS) Data
The TCDS data was used to provide a count of truck traffic at various locations. This data was integrated with the processed Geotab data to provide a more complete picture of truck activity in the study areas.

4. Final Dataset (The Dataset Used for Modeling)
The final dataset was created by merging the processed Geotab data with the relevant data from the other sources. This process involved spatially joining the Geotab and TCDS data based on location and then appending the relevant demographic and infrastructure data from CalEnviroScreen 4.0, Census, and OSM. The result is a comprehensive dataset that provides a detailed view of truck idling and parking behavior in AB 617 Disadvantaged Communities.

本研究聚焦于加利福尼亚州因邻近交通廊道、工业设施与物流中心而遭受不成比例空气污染影响的社区，重点关注与货车相关的活动——包括怠速运转、寻停车行为及停车需求，通过多维度数据集与稳健模型开展研究，所采用的技术涵盖随机森林（Random Forest）、卷积神经网络（Convolutional Neural Network）、贝叶斯岭回归（Bayesian Ridge Regression）与空间误差模型（Spatial Error Model）。主要研究结果揭示了影响货车怠速时长、寻停车耗时及停车需求的相关因素，其中重型货车怠速时长最长，寻停车难题主要集中在交通干线与货运场站周边。空间误差模型则凸显了AB 617弱势社区内货车活动、社会经济变量与空气污染之间的关联。基于上述发现，初步政策建议包括开展针对性的怠速管控宣传活动、完善货车停车设施、推广清洁燃料与技术、提升路线规划效率、收紧排放标准以及强化土地利用规划。

研究方法
本数据集提交的数据源自多个来源，各来源均为AB 617弱势社区的货车怠速与停车行为研究提供了独特视角。本次提交的数据集系对原始数据集进行精心处理与整合后所得，旨在全面呈现货车怠速与停车行为特征。

1. 吉奥塔布点火平台（Geotab Ignition Platform）数据
尽管受许可协议限制，本次提交未直接纳入吉奥塔布点火平台的数据，但该平台数据为本数据集的构建发挥了关键作用，其包含原始怠速数据、寻停车原始数据及货车停车位置原始数据。我们依托此类数据集提取了与货车怠速及停车行为相关的核心指标。研究人员通过数据中附带的地理哈希（geohash）按位置对数据进行分组，进而计算每日怠速时长与寻停车耗时，完成了吉奥塔布数据的处理与聚合。需说明的是，受许可限制，原始吉奥塔布数据未纳入本次提交。若有使用者希望使用该类数据，请直接访问吉奥塔布官网获取。

2. 加州环境筛查4.0（CalEnviroScreen 4.0）、人口普查（Census）数据与开放街道地图（OpenStreetMap，OSM）
此类数据集提供了人口统计与基础设施等背景信息，用于丰富从吉奥塔布数据集衍生的怠速与停车数据。例如，研究人员借助人口普查与加州环境筛查4.0的人口统计数据识别弱势社区，同时依托开放街道地图数据将怠速与停车行为匹配至具体地理位置。

3. 克恩县交通计数数据系统（Kern County Traffic Count Data System，TCDS）数据
该数据集用于统计各点位的货车交通流量，并与处理后的吉奥塔布数据进行整合，以更全面地呈现研究区域内的货车活动情况。

4. 最终建模数据集（本次研究使用的数据集）
最终数据集通过将处理后的吉奥塔布数据与其他来源的相关数据进行合并构建而成：首先基于地理位置对吉奥塔布数据与TCDS数据进行空间连接，随后追加来自加州环境筛查4.0、人口普查及开放街道地图的人口统计与基础设施相关数据。最终所得数据集可全面且细致地展现AB 617弱势社区内的货车怠速与停车行为特征。