This dataset contains transcripts, audio, and video of official “Arrival” and “Doorstep” speeches from European Council events, collected from the European Council Newsroom. Each entry includes metadata such as the title, date, duration, language, word/character counts, and links to audio and video sources. The speeches are mainly in European languages, with some multilingual and non-European entries.

All recordings were automatically transcribed using OpenAI’s Whisper ASR model. Items lacking audio or containing two audio channels (≈6% of recordings) were excluded to simplify processing.

Source: European Council Newsroom

Dataset Structure

Column Name	Data Type	Description
`title`	string	Title of the entry
`date`	string / date	Uploaded date on the webpage
`duration`	float	Duration in seconds
`language`	list of string	Languages spoken in speech
`transcript`	string	Full text transcript of the speech
`segments`	list[dict]	a time-aligned transcription of the media content.
`word_count`	int	Number of words in the transcript
`character_count`	int	Number of characters in the transcript
`videoTypeDescription`	string	Type of video (e.g. doorstep interview, speech)
`split`	string	Data split
`license`	string	Dataset license information
`url.audio`	string	URL to the audio
`url.video_preview`	string	URL to the preview of the video
`url.video`	string	URL to the HD video

Dataset Size

Number of records: 45522
Total duration in hours: 3673.88
Average transcript length: 3949.61 characters / 659.85 words

Language Distribution

Language	Count	Total Duration (hours)
English	23638	1754.28
German	4707	411.39
French	3967	343.80
Spanish	2705	258.43
Italian	1614	156.40
Dutch	1884	178.41
Swedish	1429	131.33
Portuguese	1300	119.43
Polish	1109	105.13
Czech	986	89.36
Danish	920	83.43
Croatian	919	79.77
Finnish	838	80.29
Slovak	800	72.32
Greek	709	46.27
Slovenian	596	48.07
Bulgarian	473	71.24
Luxembourgish	479	47.20
Hungarian	454	41.34
Romanian	350	30.13
Lithuanian	208	15.02
Maltese	177	7.28
Latvian	135	9.65
Arabic	142	10.54
Ukrainian	101	8.73
Estonian	93	6.67
Multilingual	94	7.74
Serbian	59	5.36
Russian	60	4.42
Irish	45	3.89
Norwegian	44	4.66
Georgian	44	2.99
Albanian	43	3.58
Macedonian	33	2.58
Montenegrin	27	2.16
Bosnian	24	1.99
Turkish	19	1.85
International	17	0.44
Persian	12	1.07
Armenian	10	0.51
Chinese	9	0.96
Catalan	8	0.75
Japanese	7	0.53
Belarusian	7	0.40
Moldavian	6	0.40
Indonesian	4	0.33
Basque	4	0.43
Korean	3	0.33
Vietnamese	3	0.33
Kazakh	3	0.23
Icelandic	2	0.22
Hebrew	2	0.19
Swahili	1	0.05
Hindi	1	0.10
Tajik	1	0.09
Khmer	1	0.06

Note:
For items containing multiple spoken languages, the total duration is split equally among the detected languages, as precise language-level timestamps are not available. This avoids inflating the total duration across languages.

Usage

from datasets import load_dataset

# Load the full dataset. 
ds = load_dataset("meetween/eu_council", split="eu_council")

# Load dataset in streaming mode.
dataset = load_dataset(
    "meetween/eu_council",
    split="eu_council",
    streaming=True
)
for example in dataset.take(3):
    print(example["id"], example["language"])

License

Downloads last month: 15