S. González, J. Vega, A. Murari, A. Pereira and JET EFDA contributors New automatic analysis methods allow the analysis of large amounts of data without human interaction. Tokamak machines, such as JET, are perfect candidates to apply data mining techniques in order to obtain results with a high statistical relevance. In this paper, an automatic technique to analyse the pedestal edge gradient is introduced. This technique does not require human intervention and therefore can be applied to many pulses. The pedestal edge gradient is the temperature gradient corresponding to the external transport barrier at the edge of H-mode plasmas. This gradient is quantified using the temperature profiles obtained from the ECE diagnostic. An automatic technique to locate events in plasma pulses is applied in order to locate ELMs and then, the evolution of the edge pedestal gradient is analysed during the ELMs. The degradation of the edge pedestal gradient during an ELM is quantified using the edge pedestal gradient 2ms before it as a reference of the amplitude of the gradient. This technique has been applied to a JET database containing more than a thousand pulses for a total of more than 46000 ELMs.