Data Sharing Issues: What Works and What Doesn’t

The following is a guest-post from national metadata advocate Lynda Wayne. In January, she put out a request to the NSGIC Community for input on her project to document "a set of practical measures to encourage/facilitate data sharing." This was part of an effort to assist to Dr. Tim De Troye and the Geospatial Administrators Association of South Carolina (GAASC) in developing the Process Framework for Developing Local Government Data Access Policies document (shared by Tim on this blog in January) and to add some "real world" content to a follow-up SC GIS conference presentation and discussion. Here is what she found out.

Issue: Variable Data Sharing Agreements
Data sharing agreements vary greatly in their complexity, effectiveness and requirements. When trying to share data among multiple organizations, managing differences and conflicts among agreements can be overwhelming.

What works:

• Handshakes making share Admin is on board
• Collaboratively developed, effective, standardized data sharing agreements

What doesn’t:

• Formal agreements because they typically include imprecise language that can serve as road blocks and add unnecessary overhead to the effort
• Over-involvement on the part of Administration and Legal personnel that leads to overly burdensome language in an effort to cover all the bases and possible scenarios

Issue: Missing or Out of Date Data Documentation
If data are not well-documented then they cannot be applied as fully or appropriately and competing/conflicting data sets can emerge.

What works:

• Providing metadata creation, review and management support in return for data access
• Support for a community-wide metadata training and support organization/resource

What doesn’t:

• Trying to create metadata by guessing at how the data were created

Issue: Maintaining Personal Privacy and Public Safety
Data developers have an obligation to protect the privacy and safety of their constituents.

What works:

• Utilizing the FGDC Guidelines (PDF) as to what data should be shared
• Utilize available standards (e.g. URISA/FGDC Addressing standard) that address privacy and security issues
• Generalizing data such that non-sensitive content is still available
• Public Record Laws and ‘policies from above’ that explicitly include geospatial data and establish clear guidelines as to who can access the data and how to request access
• Recognizing that very little data is truly ‘sensitive’

What doesn’t:

• Wholesale approaches that eliminate sharing of all ‘potentially’ sensitive data
• Over-involvement on the part of Administration and Legal personnel that attempt to cover all the bases

Issue: Maintaining Control of the Data
Data, like any resources, is subject to personal, professional and/or political control issues.

What works:

• Earnest dialog to identify precise concerns and brainstorm on innovative solutions
• Naming conventions and data lineage guidelines such as those published by the Louisiana Data Council (PDF)
• Data management models that allow the use of data that is maintained in your system
• Data steward / trusted source models

What doesn’t:

• Treating public data as a private resource

Issue: Data Misuse and Exploitation
Once data is in the hands of others in can be inappropriately applied or resold for profit.
What works:

• Metadata, metadata, metadata that includes valid ‘Use Constraints’, ‘Distribution Liability’ and ‘Purpose’ statements
• Clear, not ‘legalese’, explanation of license/copyright requirements and mandatory acknowledgement by the consumer
• Making data freely accessible so that consumers easily use current version rather than trying to milk the older data

What doesn’t:

• Stifling the use of data in new and innovative ways – if somebody makes a product somebody else is willing to pay for that drives the economy and enterprise.

Issue: Reduced Capacity and Infrastructure to Support Data Serving and Exchange
Most organizations have a mission to serve their own constituents and are staffed/resourced specifically to that mission.

What works:

• Money - especially if designated to build and maintain data sharing capacity – new data collection, hardware, software, training
• A simplified process that adds no burden to the data provider
• Documenting the return on investment that data sharing can yield to warrant capacity building

What doesn’t:

• One time payments for data that are not targeted to capacity growth

What say you, gentle readers? Feel free to add any thoughts in the comments section.