Investigative analysts gather data from different sources, especially from social media (SM), in order to shed light on cognitive factors that may explain criminal spatial behavior. A former research shows how tweets can be used to estimate private points of interest. Authors' aim was to demonstrate, as they extend the analysis to a wider statistical base, how social maps and Web applications could be used in investigative analysis and spatial cognition research. A total of 100 Twitter accounts with approximately 250 tweets each were submitted to common geographical techniques (measures such as Convex-Hull, Mean-Center, Median-Center, Standard-Deviation-Ellipse) in order to test the hypothesis that user areas of activity are predictable. Predictions were tested through a set of specific information: clear reference to areas of activity and clear reference about user's residence. Simple algorithms and procedures demonstrated that they could be used to predict where SM users live, giving positive results in about 4/5 cases and giving indications about their home location. In fact, all home positions were found in the Convex-Hull and most of them in the Standard-Deviation-Ellipse. Furthermore, in up to 80% of cases, houses were found within a buffer zone of 1.500 m with Median-Center as centrum (70% using Median-Center as centrum) with a minimum effectiveness threshold of 12-13 tweets. SM may help in studying people mobility and their cognition of space and, moreover, where they live, or their traveling behavior. The processing of geographical data in conjunction with the SM analysis may facilitate the construction of models describing specific behavior of people. The use of geographical information system tools and SM analysis represents an effective approach in order to acquire spatial and territorial information, referred to social relationship. The results may be used successfully in the understanding of social dynamics and for the prevention of criminal behavior.