The ionosphere is characterised by a large number of disturbances generated in response to a wide range of phenomena, including natural hazards, space weather and man-made events. Such disturbances are known as travelling ionospheric disturbances (TID). Identification of the origin of TID, especially in real or near-real-time (NRT), is an extremely difficult task, and it is one of the most interesting scientific questions. In this paper we present, for the first time, an approach for an automatic and NRT-compatible detection and recognition of the source of ionospheric disturbances in time series of total electron content (TEC) measured by the Global Navigation Satellite Systems (GNSS) method. The main idea is 1) to analyse main characteristics (such as spatio-temporal features and frequency content) of TID generated by known sources, and 2) in NRT, to rapidly examine TID’s, and, based on this information, recognize their source. Currently, our database contains TEC data series with response to earthquakes, volcanic eruptions, tornadoes, explosions, rocket launches and equatorial plasma bubbles. Our developments are important for the future assessment of natural hazards from the ionosphere, and also for NRT Space Weather nowcast and applications. Also, our work presents important information about the physical properties of TID of different origins.