A two-step process to increase successful geocoding in publicly available police stop data
A two-step process to increase successful geocoding in publicly available police stop data
复制标题
提高公开警察站数据中地理编码成功率的两步流程
DOI：
10.1080/15614263.2023.2181169
复制
发表时间：
2023
影响因子：
1.8
通讯作者：
Stewart, Connor
中科院分区：
文献类型：
--
作者：
Wallace, Danielle;Helderop, Edward;Grubesic, Anthony;Walker, Jason;Liu, Xiaoyue Cathy;Wei, Ran;Zhou, Yirong;Stewart, Connor
许多警察部门通过公开数据来响应透明度的呼吁。高质量的地址位置对于成功和准确的地理编码至关重要，尽管数据的内容和质量在数据集之间可能会有很大差异。在本研究中，我们展示了一个两步地理编码过程，该过程有助于使用传统的地理编码和Jaro-Winkler编辑距离方法将低质量的地址位置转换为地理可定位的地址，并使用来自圣地亚哥警察局的警察拦截数据。作为参考，在使用传统的地理编码方法时，只有83%的站点进行了地理编码。通过使用Jaro-Winkler编辑距离来清理站点地址字符串，我们能够对99%的站点进行地理编码。我们进一步讨论了在使用公开警务数据时，警察部门和研究人员针对数据质量相关问题的数据创建实践和解决方案。
Many police departments are meeting calls for transparency by releasing publicly accessible data. High-quality address locations are critical for successful and accurate geocoding, though the content and quality of that data can drastically vary across datasets. In this study, we showcase a two-step geocoding process that helps convert low-quality address locations into geo-locatable addresses using traditional geocoding and Jaro-Winkler edit distance methods with police stop data from the San Diego Police Department. For reference, only 83% of stops were geocoded when using traditional geocoding methods. By employing the Jaro-Winkler edit distance to clean the stop address strings, we were able to geocode 99% of stops. We further discuss data creation practices and solutions for data quality-related issues for police departments and researchers when using publicly available policing data.