Original article was published by Edmund W. Schuster on Artificial Intelligence on Medium
The Uses of Government Data in Society
What is the best way to release value to the public?
The age of the Fourth Industrial Revolution is set to cause profound changes in the field of public administration. The intense focus on machine to machine communication along with new data laws opens fresh directions. I feel the frictionless release of government data into society results in many benefits. If widespread use takes place, I expect measurable increases in multifactor productivity. With productivity growth also comes higher GDP. The effect of data adds to the state of technology, a factor of production for both goods and service. To date, the common use of government data by industry has been lacking. This is especially meaningful to note with regard to artificial intelligence and machine learning.
In America, the theory of public administration has a long history. Woodrow Wilson, a college professor and future President of the United States, put forth the seminal work to define the field. Wilson writes;
“It is the object of administrative study to discover, first, what government can properly and successfully do, and, secondly, how it can do these proper things with the utmost possible efficiency and at the least possible cost either of money or of energy.”
Since the time of Wilson’s paper, several paradigms characterize the stages of development. By the 1960s, a change in thinking happened due to the emergence of computing and management science. In response, the powers favoring a strong public administration introduced a proposal to form the National Data Center. The effort by proponents in the United States Congress did not succeed in passage for the legislation. The substance of the opposition — fears over privacy and the initial setting of a path toward an Orwellian government.
However, Arthur R. Miller notes the unanticipated consequence;
“… the Center’s demise may have been a Pyrrhic victory for the advocates of citizen privacy because federal agencies have been establishing since the mid-1960s (and are continuing the practice) their own centers for exchanging information that they feel they need sans the public scrutiny that a National Data Center likely would have received.”
Today, there are 13 agencies providing statistics at the national level as their sole purpose. In addition, 89 federal agencies generate statistics as a secondary mission. This fractionated system is an outcome of the failure to establish National Data Center in the 1960s at the dawn of the Third Industrial Revolution and the rise of electronics and computing.
Layered upon the government data system are several recent laws that call for a digital transformation to machine readable data formats. The goal is the seamless exchange of data between the federal government, the public, and industry.
In recognition of the profound changes upcoming to American society, Congress passed the Foundations for Evidence-Based Policymaking Act, which became law in January 2019. Title II contains the details the Open Government Data Act, giving guidance to federal statistical agencies. Other related laws include the Digital Accountability and Transparency Act of 2014 and the Grant Reporting Efficiency and Agreements Transparency (GREAT) of December 2019. Several comparable laws are under consideration in the U.S. House.
These laws mean that public administrators must become deeply involved in data engineering. This causes repurposing and a search for the best technologies and standards to achieve the legislative goals. The effects are across all federal agencies. Web sites like data.gov and niem.gov serve as a clearing house for information. But there is more to do. The chances exist to leverage public data in new ways, including the promotion of startups that connect government data sources to industry.
Broadly speaking, I see three areas of need to increase the effectiveness of public data:
2. Data Fusion
3. Internet-based, interoperable data and modeling network using advanced semantics
Given that over 100 federal agencies provide data to the public, discovery of the appropriate data set available on the Internet becomes a challenge. This produces the need to have advanced discovery systems in place for precise search.
If users are able to locate the appropriate data sets, the need exists to fuse data together. Since government data is in different formats the task of combing data becomes challenging.
Finally, like the flow of data, the application of computation must be smooth. This represents a long-term objective, which involves the creation of the Web of Abstractions.
In conjunction, many other issues require resolution. A sampling includes standards for transmission of metadata such as the accuracy, precision, or stationary of a time series data set. As well, federal agencies need to establish the boundary between government and business. Typical questions include; how much data manipulation should federal agencies be involved?
Those in industry cannot ignore the future role of public administration in releasing the value of government data for commerce. Like Woodrow Wilson’s breakthrough work, there needs to be new definition of public administration in the age of the Forth Industrial Revolution. The discipline must evolve to include an increased role for data engineering at all levels.
 Wilson, W., 1887. The study of administration. Political Science Quarterly,2:1.
 Henry, N., 1980. Public Administration and Public Affairs. Englewood Cliffs, NJ: Prentice Hall.
 Miller, A.R., 1969. Personal privacy in the computer age: the challenges of a new technology in an information oriented society. Michigan Law Review, 67.