宁波市奉化区智慧工地系统考勤预警管理数据

采集宁波市奉化区范围内考勤信息线下通过设备和手动录入的方式采集数据到数据库作为原始数据源。最后通过BI工具，按区域工种进行分类统计，利用折线图体现每个地区工种的考勤风险走势情况。推送给班组长、企业、劳务公司、监管部门。1）数据采集：采集宁波市奉化区范围内考勤信息线下通过设备和手动录入的方式采集数据到数据库作为原始数据源。（2）数据处理：首先对采集的考勤数据进行清洗，包括考勤数量为0的或者为null的。然后对数据在时间维度按日，项目维度按企业，区域维度按区域，进行最细级别粒度的聚合。计算得到各工地的考勤确认总数量为X，未确认考勤人数/X=未确认人数在总考勤数中的占比， 待处理数为已经采集且并未介入处理数， 已处理数为已经开始处理的数量，已完成数为已经确认考勤的数量，超时未处理数为超时未处理数指超过1天未处理数，缺勤考勤人数/X=缺勤考勤人数在总考勤数中的占比，补卡考勤人数/X=补卡考勤人数在总考勤数中的占比。（3）数据分析: 低危风险数=企业待处理数低于总比10%且超时未处理数低于5%，中危风险数=企业待处理数低于总比20%且超时未处理数低于10%，高危风险数=企业待处理数低于总比30%且超时未处理数低于15%。

This dataset takes the attendance information within the jurisdiction of Ningbo Fenghua District as the research target. The original data source is collected offline via device scanning and manual entry, then stored in the database. Subsequently, BI tools are used to conduct classified statistics by region and job type, visualize the attendance risk trend of each region and job type through line charts, and push the analysis results to team leaders, enterprises, labor service companies and regulatory authorities. The entire workflow includes three core stages:

1. Data Collection: Collect attendance data within the scope of Ningbo Fenghua District, and store the collected data in the database through two offline methods: device collection and manual entry as the original data source.
2. Data Processing: First, clean the collected attendance data by filtering out records with 0 or null attendance quantity. Then perform finest-granularity aggregation on the data across three dimensions: time (daily granularity), project (enterprise granularity) and region (regional granularity). Calculate the total confirmed attendance count for each construction site as X. The following metrics are derived:
- Proportion of unconfirmed attendance personnel: (number of unconfirmed attendance personnel) / X
- Pending processing count: records that have been collected but not yet processed
- Processed count: records that have started to be processed
- Completed count: confirmed attendance records
- Overdue unprocessed count: records that have exceeded 1 day without being processed
- Proportion of absent attendance personnel: (number of absent attendance personnel) / X
- Proportion of reissued attendance personnel: (number of personnel who applied for attendance reissue) / X
3. Data Analysis: Define risk levels based on the aggregated metrics:
- Low-risk: The number of pending processing records of the enterprise accounts for less than 10% of the total, and the number of overdue unprocessed records accounts for less than 5%
- Medium-risk: The number of pending processing records of the enterprise accounts for less than 20% of the total, and the number of overdue unprocessed records accounts for less than 10%
- High-risk: The number of pending processing records of the enterprise accounts for less than 30% of the total, and the number of overdue unprocessed records accounts for less than 15%