The emergence of autonomous vehicles has brought new challenges for pedestrian-vehicle interaction (PVI). To facilitate such interactions, many external human-machine interface (eHMI) concepts have been proposed. However, the development and evaluation of these eHMIs should be based on representative PVI scenarios in daily life, which have not been systematically constructed. The goal of this study is to: 1) Identify typical PVI scenarios and analyze their constituent elements; 2) Provide a practical method to generate PVI scenarios to support the development and evaluation of eHMIs. With a literature review and a focus group interview, we concluded four typical sets of PVI scenarios: street-crossing scenarios, starting scenarios, parking scenarios, and following scenarios. After analyzing key elements of these scenarios, we constructed, from the perspective of the human-machine-environment system, a three-dimension scenario analysis framework that consisted of pedestrian, vehicle, and environmental variables. We then used our framework to generate several common or challenging PVI scenarios where eHMIs can play a role and needs to be evaluated. Our work contributes to the development and evaluation of eHMIs in both academia and industry.