FutureGov: Drones and Open Data
By Kristine Gloria, CTSP Fellow | Permalink
A drone’s capability to capture large amounts of data – audio, sensory, geospatial and visual – serves as a promising pathway for future smart city proposals. It also has many data collection, use and retention policies that require considering data formats and structures.
Why is this worth exploring? We suggest that it opens up additional (complementary) questions about access, information sharing, security and accountability. The challenge with the personal UAS ecosystem is its black box nature comprised of proprietary software/hardware developers and third-party vendors. This leads to technical hurdles such as the development of adaptable middleware, specified application development, centralized access control, etc.
How do governments make data public?
Reviewing this through an open data lens — as our work focuses on municipal use cases –offers a more technical discussion and highlights available open source developer tools and databases. In this thought experiment, we assume a government agency prescribes to and is in the development of an open data practice. At this stage, the agency now faces the question: How do we make the data public? Additional general guidance on how to approach Open Data in government, please refer to our work: Open Data Privacy Report 2015.
Drawing from the Sunlight Foundation’s Open Data Guidelines, information should be released in “open formats” or “open standards”, and be machine-readable and machine-processable (or structured appropriately). Translation: data designated by a municipality as “shareable” should follow a data publishing standard in order to facilitate sharing and reuse by both human and machine. These formats may include XML, CSV, JSON, etc. Doing so enables access (where designated) and opportunities for more sophisticated analysis. Note that the PDF format is generally discouraged as it prevents data from being shared and reused.
Practical Guidelines for Open Data Initiatives
It seems simple enough, right? Yes and no. Learning from challenges of early open data initiatives, database managers should also consider the following: completeness, timeliness, and reliability & trustworthiness.
- Completeness refers to the entirety of a record. Again, the Sunlight Foundation suggests: “All raw information from a dataset should be released to the public, except to the extent necessary to comply with federal law regarding the release of personally identifiable information.” We add that completeness must also align with internal privacy policies. For example, one should consider whether the open data could lead to risks of re-identification.
- Timeliness is particularly important given the potential applications of UAS real-time data gathering. Take for example emergency or disaster recovery use cases. Knowing what types of data can be shared, by whom, to whom and how quickly can lead to innovative application development for utility services or aide distribution. Published data should therefore be released as quickly as possible with priority given to time-sensitive data.
- Reliability and Trustworthiness are key data qualities that highlight authority and primacy, such as the source name of specific data agencies. Through metadata provenance, we can capture and define resources, access points, derivatives, formulas, applications, etc. Examples of this include W3C’s PROV-XML schema. Identifying the source of the data, any derivatives, additions, etc., helps increase the reliability and trustworthiness of the data.
What of Linked Open Government Data?
For those closely following the open government data space, much debate has focused on the need for a standardized data format in order to link data across formats, organizations, governments etc. Advocates suggest that, linking open data may increase its utility through interoperability. This may be achieved using structured machine-processable formats, such as the Resource Description Framework (RDF). This format uses Uniform Resource Identifiers (URIs), which can be identified by reference and linked with other relevant data by subject, predicate, or object. For a deep dive on this specific format, check out the “Cookbook for Open Government Linked Data”. One strength of this approach is its capability to generate a large searchable knowledge graph. Check out the Linked Open Data Cloud for an example of all linked databases currently available. Paired with Semantic Web standards and a robust ontology, the potential for its use with drone data could be quite impactful.
No matter the data standard chosen, linked or not, incorporating a reflexive review process should also be considered. This may include some form of a dataset scoring methodology, such as the 5-Star Linked Data System or Montgomery County’s scoring system (Appendix H), in order to ensure that designated datasets comply to both internal and external standards.
Image from: MapBox Project
Hacking Drone Data
Now to the fun stuff. If you’re interested in drone data, there are a few open drone databases and toolkits available for people to use. The data ranges from GIS imaging to airport/airspace information. See MapBox as an example of work (note: this is now part of the B4UFLY smartphone app available by the FAA). Tools and datasets include:
- Open Drone Map: An Open Source Toolkit for processing Civilian Drone Imagery
- FAA Drone Registry Database: available CSV or searchable online
- Bureau of Investigative Journalism – Drone strike tallies; not machine-readable data which will require some scraping and restructuring
- Database of US Companies flying Drones (2015)
And, finally, for those interested in more operational control of their drone experience, check out these Linux based drones highlighted in 2015 by Network World.
So, will the future of drones include open data? Hopefully. Drones have already proven to to be incredibly useful as a means to surveying the environment and for search and rescue efforts. Unfortunately, drones also raise considerable concerns regarding surveillance, security and privacy. The combination of an open data practice with drones therefore requires a proactive, deliberate balancing act. Fortunately, we can and should learn from our past open data faux pas. Projects such as our own CityDrone initiative and or fellow CSTP colleagues’ “Operationalizing Privacy for Open Data Initiatives: A Guide for Cities” project serve as excellent reference points for those interested in opening up their drone data.