Dark Mode
French Avian Vocalisation Metadata
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This collection provides detailed information captured about bird recordings originating specifically from France. It is suitable for exploring the relationship between environmental changes and avian populations and is frequently used in the context of birdsong recognition challenges. The information includes location, time of recording, species identification, and technical audio details contributed by a large community of recorders.
Columns
The data structure contains 26 fields detailing each observation. Key fields include:
- id: A unique recording identifier.
- gen: The genus of the recorded bird (e.g., Sylvia or Turdus).
- sp: The specific species identified (e.g., major or mystery).
- en: The common English name of the species, frequently 'Identity unknown' or 'Great Tit'.
- rec: The name of the recorder, with Stanislas Wroza and Cedric Mroczko being among the most frequent contributors.
- cnt: The country of origin, which is uniformly France.
- loc: The specific location details, such as 'Gif-sur-Yvette, Essonne, Île-de-France'.
- lat and lng: The latitude and longitude coordinates of the recording.
- alt: The altitude of the recording location.
- type: The type of vocalisation captured (e.g., 'song' or 'call').
- url and file: Links to the Xeno-Canto recording page and the downloadable audio file.
- lic: The license URL governing the individual recording.
- q: The quality rating assigned to the recording (A and B are the most common ratings).
- date and time: The date (up to 29 July 2020) and time the recording was made.
- bird-seen: Indicates whether the bird was visually observed (often 'yes').
- playback-used: Indicates if playback was utilised (predominantly 'no').
Distribution
The dataset is distributed as a single CSV file named
birds_france.csv, which totals 15.05 MB in size. It includes metadata for approximately 20.9 thousand valid recordings. The collection features a wide variety of species, representing 302 unique genera and 481 unique species.Usage
Ideal applications include machine learning for avian sound recognition and classification. The data can be applied by researchers and data scientists to assess the rates of species migration, study species evolution across geographical regions, and conduct ecological modelling, especially concerning how air quality may affect bird populations.
Coverage
The geographical availability is restricted to France, which accounts for 100% of the records. The recording locations are widespread, with latitude coordinates ranging from approximately -49.7 to 51.1 and longitude coordinates from -158 to 167 (though most activity is concentrated within the French mainland). The recordings span a long timeframe, dating back to January 1920 and extending through July 2020. Upload dates range from November 2008 to July 2020.
License
CC0: Public Domain
Who Can Use It
- Ecologists and Biologists: For monitoring changes in species distribution and studying the effects of environmental variables.
- Data Scientists: For training and evaluating advanced sound recognition models and clustering algorithms.
- Researchers: To analyse longitudinal trends in bird populations and vocalisation patterns.
- Hobbyists: Interested in exploring acoustic biodiversity and contributing to citizen science efforts.
Dataset Name Suggestions
- Birds of France Acoustic Metadata
- Xeno-Canto France Recordings
- Avian Song Metadata (France)
Attributes
Original Data Source: French Avian Vocalisation Metadata
QUALITYLoading...