A hierarchical access setup is considered, where secondary users can (re-)use frequency bands allocated to licensed systems, provided ongoing primary communications are not overly disrupted. Since conventional spectrum sensing schemes can detect and localize 'active' sources but not 'passive' users, the number of primary receivers and their locations are generally unknown. Supposing a minimal coordination between primary and secondary systems, a novel method for unveiling areas where primary receivers are located is proposed in this paper. The primary system broadcasts short messages - here refereed to as 'interference tweets' - indicating the number of receivers that are interfered. Using these tweets, together with a grid-based discretization of the primary coverage region, the locations where receivers are likely to reside are obtained by solving a sparse linear regression problem. Subsequently, the estimated locations are used to optimize resource allocation of the secondary network operation under interference constraints.