For many agencies, what data to make open is left up to the agency’s judgment. This has worked well as agencies do a good job in understanding the public’s needs for specific datasets. Even so, as developers and citizens begin using the open datasets, there is increasing demand for specific agency datasets. The issue is how to best accommodate those requests given the constraints of agency budgets and open data support staff.
Recently, the Idea Lab at the Department of Health and Human Services (HHS) turned to lean startup methodology to develop Demand-Driven Open Data (DDOD). According to HHS, DDOD requests are converted to use cases. A use case is more than just a request for a dataset; the use case also describes who will be using the data, how the data will be used and the value of fulfilling the data request. Use cases are documented and made publicly available so other potential requestors can determine if an existing use case will meet their needs.
Along with the immediate benefits of offering HHS datasets that are in the most demand, there are long-term benefits of creating a robust community and tools around HHS open data. HHS plans to offer improvements such as “ontological tagging,” “extended dataset-level metadata,” “full data dictionaries,” and “defined cross-dataset relationships.” What does all that mean? A deeper understanding of the HHS datasets, how the datasets are related to each other and, most importantly, how to use the datasets most effectively.
In my opinion, the greater benefit is the community of developers, researchers, academics, entrepreneurs and citizens that will be built using the use cases information. As HHS dataset users observe what other users are doing with the data, this will encourage collaboration and, possibly, crowdsourced innovations based on HHS datasets. Like the agricultural big data ecosystem I wrote about in a previous column, there are numerous producers and consumers of health data. The HHS DDOD platform can be a great way to map the health data ecosystem while spurring more beneficial uses of federal health open data.
I hope that the DDOD approach is adopted by other agencies and that a federal government-wide DDOD platform is created. I’ve written before about the benefits of combining disparate datasets to create new data products. Imagine the benefits of combining health data with educational data and labor statistics data. HHS has created a great tool for uncovering the hidden value in federal open data.DDOD was made possible by HHS’s Entrepreneur-in-Residence program.