-Downloading all JRE podcasts and scrap metadata.
-Silence and noise notebook preprocess the audio signal so that an analysis of speakers' interventions and silence durations is done, results are in df_silence.csv and df_noise.csv in data folder. Tuned version is implemented using a Gaussian Mixture Model.
-File turns.csv inside data folder is used for annotating speakers' turns in oder to generae a dataset for a classifier.