(Header image by Matt Cartney, Flickr Creative Commons License)
The adoption of web technologies offers national, regional, and local governments around the world a new way to disseminate information and promote transparency. Important information such as budgets, government salaries, detailed government expenditures, and meeting records can be conveniently posted by government officials and easily accessed by community members and other stakeholders. In recent years, civil society groups have developed transparency standards for local and national governments and have sought to “grade” government websites on the degree to which they meet these standards.
Existing civil society efforts to identify governments who fail to meet these transparency standards have been spotty and episodic due to the time and resources required. To address this problem, a robust and scalable methodology is needed to evaluate government websites. At MIT GOV/LAB, we are creating a “big data framework” for evaluating digital governance at all levels of government, with a focus on local governments. Using an approach driven by machine learning, we hope to effectively classify tens of thousands of government websites in a diverse range of countries. The end goal of this evaluation is to provide data for government administrators, civil society, and the press and to ultimately establish a data-driven standard for government transparency online.