Driver Insurance Premium Calculation using Advanced Driver Assistance Systems and Contextual Information: dataset

The dataset comprises driving data collected from an Irish fleet of 354 commercial drivers using light-good vehicles in a driving monitoring campaign. As part of their daily job activities, the drivers performed, on average, five daily trips, covering around 143 km. The data collection occurred in the Republic of Ireland between April 2021 and April 2022, encompassing 8,142,896 km from 287,511 trips, where drivers were monitored for 35 weeks on average. The fleet was monitored with telematics tracking devices and warning-based ADAS that triggered alarms about distraction-related events. With this information, all drivers received feedback about their driving patterns and attended quarterly coaching sessions to meet the fleet’s standards regarding road safety.

Driving behaviour patterns are often constrained by the driving environment. Thus, geolocation data was used to retrieve the context of all daily trips. This contextualization was performed using the Motion-S contextualization service based on HERE Technologies's Platform Data Extension. As a result, this dataset contains weekly aggregated driving data including behavioural patterns, distraction-related events through ADAS warnings, and the driving context.

The dataset also contains information about two years of the fleet claims history (2020 and 2021). In total, 62 at-fault claims were recorded, giving a mean cost per claim of €2,899, where 50 drivers had one claim and six drivers had two claims.

Files:
dataset_description.md (txt file with description and variables)
dataset_description.ipynb (Python notebook)
weekly_driving_profiles.csv (data)

本数据集采集自一项驾驶监测项目，涉及爱尔兰共和国354名驾驶轻型载货汽车的商业驾驶员。驾驶员在日常工作中平均每日完成5趟行程，累计行驶里程约143公里。数据采集工作于2021年4月至2022年4月间在爱尔兰共和国开展，累计覆盖287511趟行程，总行驶里程达8142896公里，驾驶员平均被监测时长为35周。该车队通过远程信息处理（telematics）追踪设备与预警式高级驾驶辅助系统（ADAS, Advanced Driver Assistance Systems）进行监测，该系统可针对分心类驾驶事件触发告警。基于采集到的监测数据，所有驾驶员均可获得自身驾驶行为模式的反馈报告，并参与季度培训课程，以符合车队制定的道路安全标准。

驾驶行为模式通常受驾驶环境制约，因此本数据集借助地理位置数据还原所有日常行程的行驶场景。该场景还原工作依托基于HERE技术公司（HERE Technologies）平台数据扩展功能的Motion-S场景还原服务完成。综上，本数据集包含按周聚合的驾驶数据，涵盖驾驶行为模式、通过ADAS预警记录的分心类事件，以及驾驶场景信息。

本数据集同时收录该车队2020年至2021年共两年的车险理赔历史数据。总计记录有62起有责理赔案件，单起理赔平均赔付金额为2899欧元，其中50名驾驶员各涉及1起理赔，6名驾驶员各涉及2起理赔。

数据集文件包括：dataset_description.md（含数据集说明与变量信息的文本文件）、dataset_description.ipynb（Python笔记本文件）、weekly_driving_profiles.csv（数据文件）