GB biodiversity and cultural ecosystem services paper dataset
Lists
Only you see this
Abstract
Scripts and data associated with the publication "<strong><em>Social media and deep learning reveal specific cultural preferences for biodiversity</em></strong>". This repository contains a series of scripts (listed below) and key data including training weights and predictions from the beta=0.01 species interaction model and predictions from the species classification model. The prediction files contain Flickr and iNaturalist ids which can be linked to metadata accessible through the platform APIs. <strong>To train the human-species interaction model:</strong> 1. sample_grid.R (generate sample grid over Europe) (2.) For Flickr, we extracted metadata in Europe using the API and the sample grid coordinates* 3. inat_extract.R (extract metadata using iNaturalist API) 4. sample_inat_flickr.R (sample Flickr and iNaturalist metadata for model training, then split into train/val/test) 5. flickr_download.py (download Flickr images in sample boxes) 6. inat_download.py (download iNaturalist images in sample boxes) 7. diff_model_train.py (train the model using different beta coefficients) 8. diff_resnet.py (the model) 9. diff_data_loader.py (the data loader) <strong>To download Flickr and iNaturalist data for Great Britain:</strong> (10.) For Flickr, we used metadata and images downloaded in a previous study using the Flickr API* 11. inat_obs_gb.R (download iNaturalist metadata and add extra taxonomic information). <strong>To apply the iNaturalist species classification model:</strong> 12. inat_apply_gb.py (Flickr dataset) 13. inat_apply_gb_obs.py (iNaturalist dataset) 14. inception.py (iNaturalist model) <strong>To test, apply and integrate the predictions of the human-species interaction and species classification models:</strong> 15. predict_test.py (apply the model to the test dataset) 16. predict_gb.py (apply the model on the Flickr image dataset for Britain) 17. data_loader.py (data loader) 18. gb_metadata.R (attach species interaction model predictions to gb flickr metadata) 19. attach_species.R (attach species classification predictions to gb flickr and inaturalist metadata) 20. nat_model_stats.R (produce model accuracy results) <strong>To conduct analysis:</strong> 21. naturalist_activity.R (genus and spatial analysis of interactions on Flickr and iNaturalist) 22. biodiversity_comp.R (comparison with bird species richness and abundance) 23. species_time (comparison with threatened migratory birds) <strong>Functions:</strong> 24. functions.R (all functions) *additional scripts and data can be found here: https://doi.org/10.5281/zenodo.5534028
BibTeX
@inproceedings{havinga2023gb,
author = {Ilan Havinga and Diego Marcos and P.W. Bogaart and Dario Massimino and Lars Hein and Devis Tuia},
title = {{GB biodiversity and cultural ecosystem services paper dataset}},
booktitle = {Zenodo},
year = {2023},
doi = {10.5281/zenodo.7675393},
}Loading related content...