The Problem: Data Privacy in Data Sharing Initiatives
Cities and transportation agencies are well aware of the benefits that data can bring to operations and planning. More streamlined passenger journeys, prosperous communities, optimized maintenance routines, operational performance improvements, and an understanding of demand and traveller flows are just a few of the benefits. As the transportation landscape widens beyond public agencies to include private providers (e.g., TNCs), micro mobility, and more, the stakeholders for those datasets are ever growing.
Data sharing is becoming more important, but also more complex.
A critical example of the difficulties in mobility data sharing is illustrated in the clash between Uber and the LA Department of Transportation over sharing scooter data. LA DOT defined the Mobility Data Standard (MDS) as a data sharing format for cities to collect and share information for micromobility assets. L.A. required that Uber provide scooter usage data in MDS format, as a condition for Uber’s scooter operations in Los Angeles. Uber pushed back, arguing that requiring sharing of mobility data violated the California Electronic Communications Privacy Act. Uber now has an open lawsuit against LA DOT.
The crux of this controversy is data privacy.
Cities, transit agencies, and transportation agencies are challenged with balancing user privacy concerns with the benefits of collecting, analyzing, and sharing the growing data on their systems, networks, and streets. Sharing data within the transportation community can help optimize schedules and availability across mobility options. Shared data in the hands of the public can spur innovation in the economy by uncovering new capabilities. The importance of data privacy becomes even more pronounced in the European context with General Data Protection Regulation (GDPR), and ever evolving local and regional policies. The heart of the issue is to enable data sharing securely to accommodate data privacy while still providing data rich enough to provide valuable information and enable analysis.
Cubic faced this issue first-hand.
One of our North American customer’s Automated Fare Collection (AFC) system (we can’t say whose) contains an immense amount of valuable data that they desired to make available to different user groups both within and outside their organization. However, they were limited in their ability to share the data their system produces due to privacy concerns and data sharing restrictions.
The agency had to decide between two options, neither providing the desired outcome:
- Make as much of the high value, raw data available, including sensitive and PII data, knowing it can only be shared with a small audience.
- Strip away much of the data to allow it to be shared with a much larger audience, but in the process lose much of the analytical value as well.
To solve this challenge, the Cubic Global Analytics team worked closely with the agency stakeholders to understand their information needs, then configured and deployed Data Management and Analytics Platform (DMAP)’s De-Identification Engine. This solution ‘depersonalizes’ PII and highly sensitive data, making it safe for consumption by different user groups with varying levels of data access authorization.
The analytics ready data models produced by the system strike the right balance between data privacy and security, while also retaining the valuable insights needed by the range of stakeholders:
- Academic and research partner institutions
- Public consumption
DMAP also provides different methods and channels for accessing data, ensuring that the right data reaches the right stakeholders at the right time. Authorized agency ‘power users’ can sign in to DMAP’s Analytics Content Portal to explore data, and create, save, and load their own PowerBI dashboards built with depersonalized data. APIs with tiered access based on stakeholder groups make data available to be ingested into other applications and downstream systems, including those that make it safely available to the public.
The Solution: De-Identification Engine
Public and private entities generate and have access to a lot of data that can drive efficiencies and insights when shared internally, and externally with public-private partnerships and open data initiatives. Anonymized travel data can be shared while still masking any individual’s travel patterns. This data can then be used, for example, to make passenger travel information easily accessible to the public, via websites and APIs. Open data initiatives and university partnerships can both provide additional value using this data, in the form of digestible information, insights based on the data, and third-party apps made accessible to the public and the transit agencies.
These entities also have a responsibility to protect user data. That is why Cubic designed the De-Identification Engine, which:
- Transforms sensitive and PII data into an analytics-ready dataset
- Makes data safe for internal and external audiences, including open APIs
- Applies de-personalization techniques based on the needs of use cases and data sets
- Includes removal, hashing, generalization, aggregation, and more
- Simplifies data automation process with embedded privacy controls
The De-Identification Engine is configured to meet individual customer needs, preferences, and data sources, and carefully balances the analytical value that can be derived from the data with the amount of depersonalization that is applied. Each data source field is analyzed to determine the specific depersonalization technique that can be applied. Techniques include: removal, hashing, generalization, and aggregation. After the data has been transformed, it is then prepared for consumption and can be made available via a secure content portal, APIs, or simply delivered to another secure location.
Looking to the Future
Data concerns are not going away. Data sharing and open data initiatives will likely become even more important as the transportation industry grows more interdependent among citizens, public agencies, cities, and private companies. In an internal context, de-identification of data allows data to be shared across an organization, allowing all users to access insights, and a common picture of demand and service performance across the network. This allows marketing, planners, and operations teams within transit agencies to access the same secure data when doing short term and long term planning. This also enables data sharing between agencies and transit operators which have adjacent service areas, and allows them to optimize timetables and typical transfer points. In an external context, de-identification allows for safe data sharing across different public, private, and community stakeholders, and lays the foundation for collaboration, interoperability, and common understanding, while putting privacy first. De-identification is not only applicable for cities and public agencies, but companies across industries can also use de-identification techniques to make sensitive data safe for internal and external consumption, particularly as it relates to cooperating with broader community initiatives like spurring economic development.