An essential task in critical infrastructure protection is the assessment of critical infrastructure vulnerabilities. The use of scenario sets is widely regarded as the best form for such assessments. Unfortunately, the construction of scenario sets is hindered by a lack in the public domain of critical infrastructure information as such information is commonly confidential, proprietary, or business sensitive. At the same time, there is a wealth of municipal data in the public domain that is pertinent to critical infrastructures. However, to date, there are no reported studies on how to extract only the most relevant CI information from these municipal sources, nor does a methodology exist that guides the practice of CI information mining on municipal data sets. This problem is particularly challenging as these data sets are typically voluminous, heterogeneous, and even entrapping. In this chapter, we propose a knowledge-driven methodology that facilitates the extraction of CI information from public domain, that is, open source, municipal data sets. Under this methodology, pieces of deep, though usually tacit, knowledge acquired from CI domain experts are employed as keys to decipher the massive sets of municipal data and extract the relevant CI information. The proposed methodology was tested successfully on a municipality in the Southeastern United States. The methodology is considered a viable choice for CIP professionals in their efforts to gather CI information for scenario composition and vulnerability assessment.