Recommendations for Better Citywide Compliance

In the 2016 E&V report, MODA wrote that empowering ODCs was fundamental to improving compliance with the Open Data Law. The 2017 E&V report shed additional light on the challenges ODCs face – namely, that different ODCs have unique needs according to their role and the type of operations and services delivered by their agencies. The 2018 process reinforced 2016 and 2017 recommendations listed in Appendix F, and also raised the following five items.

Agencies should embed data analysts in their Open Data programs to support in the preparation of datasets for publication.

Data analysts are often familiar with business cases for data and can advise on how to best format data releases to be useful for public users.

The Open Data team should provide guidance on whether and how to disclose specific common data types and elements to Open Data.

The Open Data team should work with the City’s Chief Privacy Officer, the Law Department, and DoITT to provide agencies guidance on whether and how to release specific categories of data, such as information collected from sensors and employment record numbers. These guidelines should be prescriptive and reflect best practices for data classification and information disclosure through Freedom of Information Law (FOIL) requests.

The Open Data team should develop guidance and provide oversight on publishing inter-agency and integrated data products.

The agencies that completed the E&V process commented that they share datasets with other agencies, including the Mayor’s Office, for inter-agency initiatives. There may be more utility in coordinating releases of integrated datasets and data tables that share a common identifier as collections, rather than by each agency individually.

Agencies should work with DoITT to explore technology enhancements that improve data discovery, compliance reporting, and stakeholder management.

The E&V dataset questionnaire provided a structured way to evaluate where data existed across agencies, but ODCs said the spreadsheet format was cumbersome and that the process was time consuming for staff. The Open Data team should explore a data and metadata management solution that connects directly to agency data systems, such as data catalog systems or logical data warehouses. This could allow agencies to track datasets by IT system and manage compliance as new data is created, rather than after-the-fact during annual reporting. This would greatly benefit large agencies such as DOT that have hundreds of datasets to track across dozens of agency divisions and data owners. It would also help data discovery shift from a dataset-orientation to a systems-orientation, allowing for more holistic evaluation of dataset releases.

Agencies need a way to track and manage their data inventories, including performance and usage metrics.

Specific improvements include:

1. Agencies should be able to make direct edits to metadata for their Open Data inventories.

2. Agencies should have a way to know when dataset automations fail. Currently, agencies do not know if an automation fails unless a user complains about stale data. A more proactive quality assurance approach to inventory management would allow agencies to respond soon after a problem occurs.

3. Agencies should have a way to understand usage of their inventory and to understand the status existing data feeds.