I suppose it would involve understanding how people flow through the area. You wouldn't want to place a market or food truck on the corner where nobody goes, nor would you want to place it directly in front of a train station where the flow of traffic might become clogged. The geospatial data allows us to identify specific areas in a city, and the observation allows us to identify specific locations in the area (kinda like drilling down into the data). Understanding (by observing) the flow of people will provide valuable information to deploying an effective intervention.
I really like the idea of leveraging location-data to identify gathering areas. A great way to follow-on to this would be to perform observational research at the identified locations to understand how people interact there and understand where there are opportunities for trucks and markets to "put down roots."