Description:
The City of Philadelphia has migrated metadata for this data set to the City-wide Metadata Catalog. For any questions about the migration and administration of the Metadata Catalog, please contact maps@phila.gov.
To view the metadata for this layer, please visit: https://metadata.phila.gov/#home/datasetdetails/5543868920583086178c4f8e/representationdetails/570e7621c03327dc14f4b68d/
Below is information copied automatically from metadata.phila.gov:
Part 1 & Part 2 Crime Incidents from the Police Department's INCT system with generalized UCR codes and addresses rounded to the hundred block. These counts may not coincide exactly with data that is submitted to the Uniformed Crime Reporting (UCR) system.
Development Process:
The Philadelphia Police Department (PPD) uses a legacy mainframe system as its records management system (called INCT/Police System). Being a legacy system, INCT cannot directly transfer data to the Open Data Philly data server. The current set up is for the INCT to do a massive data dump around 3a in the morning (this extra 3 hours allow for data entry for the incidents occurring after midnight). The data is then ingested into a modern Police GIS data server that supports a variety of applications. This ingestion process completes by 6 am (this takes 3 hours, as it is not just adding records from the night prior; it also updates historical records as determined by investigators). Finally, the Police GIS data server transfers this updated data to Open Data Philly by ~10am.
***Of note - There has been a transition of our data systems early this year (2020). In addition to switching data sources (from on-premise servers to a cloud-based system), this effort also included a comprehensive review of the data being uploaded. In the previous version, we were including "stolen auto recovered" as part of the stolen auto data, as this was one of the metrics that the department keeps track of internally (also in compliance with FBI UCR). However, for the public, it has come to our attention that "recovered auto" as a crime incident was confusing (even though recovered auto could be flagged/removed/excluded using text_general_code). Thus, we removed recovered autos as incidents from Open Data, which in turn resulted in the reduction in the overall incident count. Please note that the original stolen incident is still in ODP (the subsequent recovery of the vehicle is the one removed from the Open Data incidents); so, any analysis of stolen auto patterns should not be affected by this change.