Creating an Informed Public

As cities and towns have started to progressively use more data to inform the decisions they make about services, the quality of life of citizens, and economic development, the divide between the people affected by these decisions and the data used to make them may become much wider. This sort of abstraction moves away from the principles of open governance and democratic community engagement that enables the public to give their valuable input into the way their municipality functions.

Applications and Solutions: Open Data

Open Data is an approach to sharing the data that municipalities use to make decisions which seeks to share that data with the public in new and accessible ways. This can include population statistics, unemployment rates, the municipality’s operating budget, information on public and private transportation – basically, any aspect of life in the municipality.

Technologies

Open Data Registries – Municipalities may participate in a cities/township registry administered by a third party service. This registry uses data provided by the municipality to allow for accessible comparisons between the infrastructure, services, and populations of municipalities.

Data Portals and Databases – Other Canadian cities have created their own data portals which host open data on Transportation, Health & Safety, Business & Finance, Infrastructure, and other areas.

Data Request Systems – Provides data on an individualized basis by request. Requests may be provided to the requestor only or may be added to a communal data portal / database.

Case Studies – Case studies of how citizens of the municipality are using existing open data initiatives can help encourage others to use open data in their own projects and initiatives.

Managing Liability Issues

Privacy

Issues.

⚠️ While the objectives behind open data initiatives are laudable, of concern is if individuals may be able to be identified through the use of available data.

Managing Issues.

✅ Conduct annual audits and/or risk assessments. Doing so could raise issues of privacy that may not be examined otherwise.

✅ Reassess data as new facts emerge. Risk analysis should not be a static process. If new information emerges and is to be added to a dataset, the assessment should again be conducted to avoid disclosing personal information inadvertently. Legacy data should undergo evaluation as well.

✅ Anonymize data. Involved parties should likely not be identified beyond their approximate age and gender, if they are identified at all.

✅ De-identify at the source. Extraneous information should not be gathered when data is being obtained for use in datasets if it is unnecessary.

✅ Do not publish data that could identify individuals. If data cannot be adequately scrubbed then it should not be published, or the request for the data should not be fulfilled on privacy grounds.

✅ Use limited-access or controlled-access models for more sensitive datasets. Data enclaves, contractual safeguards, or tiered access models can be used to limit access to potentially sensitive data to organizations that follow good privacy and security practices.

✅ Have sensitive data be reviewed by an internal review board. A diverse board that includes a variety of experts and representatives will help ensure the risks of disclosure are measured in an accurate, balanced manner. Community stakeholders should also be included.

✅ Provide notice and allow for opt-out. If data provided by an individual may be disclosed through a municipality’s open data programs or platforms, they should be informed of this disclosure and its nature and scope. Objections should factor greatly into risk assessments and given heavy consideration.

✅ Redact if necessary. Data can be redacted to exclude identifying features.

✅ Follow good privacy practices.

Inclusion: Accuracy, Equity, and Fairness

Issues.

⚠️ Because open data is simply a manner of disclosure, many issues that underlie data collection are also inherent in open data. This includes problems of accuracy, equity, and fairness. Inaccurate data may have little impact, such as if the incorrect make and model of some vehicles in a fleet is disclosed. In other cases, inaccurate data may result in poor or inefficient decision-making, unethical or illegal data uses, or discriminatory outcomes. Predictive policing and criminal sentencing, for example, often show racial bias and therefore the data generated from these areas may disproportionately and unfairly represent certain groups of people.

Managing Issues.

✅ Follow strict privacy practices. “Clean” data that does not identify individuals addresses many adjacent issues of equality and fairness.

✅ Create correction and complaints systems. Municipalities should create clear and simple methods for the public to identify and correct inaccurate data.

✅ Reassess data as new facts emerge. Data should be amended or corrected as necessary.

✅ Establish review schedules for published data. All data, agreements, and licenses should be periodically reviewed in order to assess if it needs to be updated or amended.

✅ Follow inclusive practices.