Geospatial cognition to sophisticated urban space is an essential capability to make various location-based decisions for our daily urban lives. To adapt ourselves to an unfamiliar or ever-evolving city, we need to develop urban cognition which usually requires lots of experience taking time and efforts. Moreover, it must be a tiresome work to find and ask knowledgeable people who have enough experience to a local area to learn what we would like to know on the spot. In order to collect and utilize crowd’s urban cognition probably obtained from living experience, we attempt to explore geospatial cognition of people through common experience from location-based social networks which can be regarded as a fruitful source of crowd-experienced local information. In particular, we propose a method to extract crowd’s movements as a direct and useful hint to know common urban cognition and measure relative socio-cognitive distances between urban clusters. In order to intuitively and simply represent cognitive urban space, we generate a socio-cognitive map by projecting the cognitive relationship into a simplified two-dimensional Euclidean space by way of MDS (Multi-Dimensional Scaling). In the experiment, we show a socio-cognitive map significantly representing cognitive proximity among urban clusters in terms of crowd’s movements from massive lifelogs over Twitter. We also provide a practical use case for nearest neighbor areas search on the cognitive map.