Continuous global monitoring of lightning has been important in recent years to study a possible relationship with global warming. Consequently, several networks to detect lightning have been installed at various spatial scales (regional and global). One of these networks is the World Wide Lightning Location Network (WWLLN), which has been monitoring lightning since 2003. It is also important to monitor the occurrence of thunderstorms and this can be roughly archived by detecting lightning clusters. In this work, we propose a lightning grouping methodology, in order to estimate the global number of thunderstorms. Our methodology consists of grouping the WWLLN data into a density matrix with a spatial resolution of 0.1° x 0.1° (1 pixel) and a temporal resolution of one hour, then the algorithm identifies the pixels with lightning and groups these pixels with the adjacent pixels to form the thunderstorm. Then, we calculate monthly, seasonal and annual averaged daily curves of the number of thunderstorms. The data set under study includes years 2012 and 2013. Our methodology is validated by calculating the linear correlation coefficient (R) between the annual daily thunderstorm curve and the “universal” Carnegie curve (R = 0.97) and with the Vostok electric field measurements (R = 0.98). Additionally, we found higher correlation in September–October–November months (R = 0.98) compared with June–July–August months (R = 0.75) for Carnegie and for Vostok station (R=0.99 and R=0.88, respectively).