The Dataset

All information on this site is based on a dataset of crawled podcast feeds. Podcasts are added to the dataset by crawling the iTunes podcast directory and fyyd. Additionally feeds can be submitted manually. Feeds are crawled according to the update interval specified in the feed, or every 24 hours if no interval is specified.

The dataset contains 261985 podcasts with 36741920 episodes. 3990141 episodes have been analysed for their audio properties.

Due to the selection process, the dataset is biased towards high ranking podcasts. Using fyyd as a source for new podcasts, introduced a bias towards German podcasts. The podcast language distribution is shown below.

Language	Podcasts	Percentage
English	128666	49.1%
Spanish	27048	10.3%
French	21748	8.3%
German	20719	7.9%
Dutch	11715	4.5%
Portuguese	11132	4.2%
Swedish	7566	2.9%
None	6402	2.4%
Chinese	4291	1.6%
Russian	3771	1.4%
Italian	2387	0.9%
Arabic	2264	0.9%
Persian	1406	0.5%
Polish	1149	0.4%
Danish	1080	0.4%
Japanese	1038	0.4%
Turkish	891	0.3%
Catalan	882	0.3%
Norwegian	700	0.3%
Ukrainian	604	0.2%
Other	6504	2.5%

The following charts shows the number of analysed episodes in relation to the published episodes.