Given the limitations of traditional methods of data collection and the increased use of smartphones, there is growing attention given to using smartphone apps for activity-travel surveys. Smartphones, through their location-logging capability, allow for the collection of high-quality data on the travel patterns of individuals over multiple days while minimizing the burden on those being monitored. This paper presents the results of an investigation into the potential and limitations of smartphone apps as passenger travel survey instruments. It evaluates the accuracy and performance of various smartphone apps using properly recorded ‘ground truth’ data. Through an open and global invitation to travel survey app and trace processing suite developers, a total of 17 apps were recruited for testing. A controlled experiment was devised, and the accuracy of the apps evaluated based on their ability to reproduce ground truth trip information. Further, the performance of the apps in terms of battery drain was also quantified and evaluated. Results indicate that while accuracy in terms of the trip ends/starts is reasonably high in most cases, mode inference accuracy varied significantly, with a maximum 65–75% accuracy achieved. As such, until significant improvements in mode inference algorithms arise, purely passive location-logging smartphone apps cannot serve as full-fledged automated travel survey instruments. While this may seem problematic, with minor input from respondents regarding regularly visited locations and modes used, as well as specific test case tuning and use of external data such as General Transit Feed Specification, there is an excellent potential to significantly reduce overall response burden and allow for high quality multi-day travel diary data to be collected. Implications of our findings for app design are discussed.

中文翻译：

---

我们到了吗？评估智能手机应用程序作为活动旅行调查的成熟工具

鉴于传统数据收集方法的局限性以及智能手机使用的增加，人们越来越关注使用智能手机应用程序进行活动旅行调查。智能手机通过其位置记录功能，可以收集有关个人多日旅行模式的高质量数据，同时最大限度地减少被监控者的负担。本文介绍了对智能手机应用程序作为乘客旅行调查工具的潜力和局限性的调查结果。它使用正确记录的“基本事实”数据评估各种智能手机应用程序的准确性和性能。通过对旅行调查应用程序和跟踪处理套件开发人员的公开和全球邀请，总共招募了 17 个应用程序进行测试。设计了一个对照实验，以及根据应用程序重现真实旅行信息的能力评估应用程序的准确性。此外，还对应用程序在电池消耗方面的性能进行了量化和评估。结果表明，虽然在大多数情况下，行程结束/开始的准确度相当高，但模式推断的准确度变化很大，最高达到 65-75% 的准确度。因此，在模式推断算法出现重大改进之前，纯粹的被动位置记录智能手机应用程序不能作为成熟的自动旅行调查工具。虽然这似乎有问题，但受访者对定期访问的位置和使用的模式以及特定的测试用例调整和外部数据的使用（如通用交通馈送规范）的意见很少，在显着减少总体响应负担和允许收集高质量的多日旅行日记数据方面具有极好的潜力。讨论了我们的发现对应用程序设计的影响。