In this article, a framework for planning sidewalk-wise paths in data-limited pedestrian environments is presented by visually recognizing city blocks in 2D digital maps (e.g., Google Maps, and OpenStreet Maps) using contour detection, and by then applying graph theory to infer a pedestrian path from start to finish. Two main problems have been identified; first, several locations worldwide (e.g., suburban / rural areas) lack recorded data on street crossings and pedestrian walkways. Second, the continuous process of recording maps (i.e., digital cartography) is, to our current knowledge, manual and has not yet been fully automated in practice. Both issues contribute toward a scaling problem, in which the continuous monitoring and recording of such data at a global scale becomes time and effort consuming. As a result, the purpose of this framework is to produce path plans that do not depend on pre-recorded (e.g., using simultaneous localization and mapping (SLAM)) or data-rich pedestrian maps, thus facilitating navigation for mobile robots and people with visual impairment. Assuming that all roads are crossable, the framework was able to produce pedestrian paths for most locations where data on sidewalks and street crossings were indeed limited at 75% accuracy in our test-set, but certain challenges still remain to attain higher accuracy and to match real-world settings. Additionally, we describe certain works in the literature that describe how to utilize such path plans effectively.