Data from: Behind the lenses: Biases in the contribution of wildlife photography to biodiversity representation

Ruíz-Villar, Héctor; Morales González, Ana1; Morant, Jon 2

Published Jul 16, 2026 on Dryad. https://doi.org/10.5061/dryad.qfttdz0s5

Data files

Jul 16, 2026 version files 311.84 KB

README.md

7.75 KB
WPOTY_Database_Ruiz_et_al_dryad.csv

304.09 KB

Abstract

Nature-related visual media has a great impact on today's society by engaging the public in conservation problems and promoting pro-environmental behaviors. Although major attention has been paid to how some types of visual media (e.g., documentaries) offer unrealistic portrayals of the natural world, biases on biodiversity representation by wildlife photography remain unexplored. In the present study, we assessed for biases in wildlife photography (understood as how biodiversity representation in photos differs from the real world) at spatial, temporal, taxonomic, conservation status, and selection criteria scales, and modeled the factors influencing the probability of portrayed organisms winning a wildlife photography contest by using data on 1333 pictures featured in the Wildlife Photographer of the Year, one of the most popular wildlife photography contests worldwide. The representation of biomes mostly coincides with it extension in the planet. However, we detected an overrepresentation of temperate (broadleaf and conifer), Mediterranean and tropical forests. We detected a positive change over time in representing historically neglected taxa such as insects. We also detected an increase in representation of Mangroves, Marine Ecosystems, Tundra and temperate forests and grasslands. Mammals and birds were overrepresented in photos while insects or plants were underrepresented, and so were species listed as “Least Concern”. The top 10 ranking species mostly included charismatic carnivore species. Our results showed that the jury’s choice offered a more diverse representation of biodiversity than the people’s choice, and the winning photographs showcased fewer taxonomic groups than the non-winning pictures. Realm, domain, and colorfulness influenced the probability of an organism’s picture being awarded winner, but the variability explained by our model reflects that there are a large number of unexplored determinants (e.g., socio-economical or technical). Our research detected a trend towards a more balanced representation of the natural world in wildlife photography, although biases are yet large, which may wrongly influence people's perception of the current status of species and habitats they encompass. Our results highlight a need to evenly represent species and ecosystems to increase public awareness, which requires providing data on species identity and conservation status to increase public knowledge. Finally, we underscore the need to report compliance with ethical guidelines when photographing wildlife.

Dataset DOI: 10.5061/dryad.qfttdz0s5

Description of the data and file structure

Héctor Ruiz-Villar, Ana Morales González and Jon Morant. 2024. "Behind the Lenses: Biases in the Contribution of Wildlife Photography to Biodiversity Representation."

This dataset accompanies an analysis of taxonomic, geographic, and conservation-status biases in the species represented in the Wildlife Photographer of the Year (WPOTY) contest. It contains the main data file WPOTY_Database_Ruiz_et_al_dryad.csv, with 1,430 entries extracted from 1,333 photographs sampled from the WPOTY contest webpage, together with a supporting lookup file. Each row corresponds to a single photograph entry, and each column records a parameter evaluated for that entry (species identity and taxonomy, generalized contest metadata, conservation status, and geographic/biogeographic classification).

This version of the dataset has been generalized to reduce the potential for identifying individual photographs (and the photographers associated with them) and for disclosing sensitive species-occurrence information. See the "Data de-identification and sensitive-data statement" section below for the specific steps applied.

File format

Both files are UTF-8 encoded plain-text CSV.
The main data file WPOTY_Database_Ruiz_et_al_dryad.csv is semicolon (;) delimited (several fields contain internal commas).
Missing / not-available values are represented by empty cells.

Files and variables

File: `WPOTY_Database_Ruiz_et_al_dryad.csv`

This file contains all entries (n = 1,430) extracted from the photographs sampled from the WPOTY contest webpage (n = 1,333), alongside each parameter evaluated. It has 20 columns:

Column A — year: Year of the WPOTY contest edition, generalized into 3-year bins (2010-2012, 2013-2015, 2016-2018, 2019-2021, 2022-2024).
Column B — com_name: Species common name.
Column C — sci_name: Species scientific name.
Column D — specified: Whether the species name was explicitly indicated on the contest webpage (Yes/No).
Column E — category_group: Generalized contest category. The original fine-grained contest categories (including individually named and special awards) have been collapsed into 12 broad groups to reduce the specificity of contest/award information. The full mapping from original category to group is provided in category_grouping_lookup.csv. Groups: Behaviour; Animal portraits & environment; Animal taxonomic category; Plants & fungi; Underwater & aquatic; Habitats & environments; Urban wildlife; Creative & artistic; Photojournalism & conservation; Young photographer; Portfolio award; Special award.
Column F — Continent: Continent at which each photograph was taken.
Column G — kingdom_name: Kingdom name for each species.
Column H — phylum_name: Phylum name.
Column I — class_name: Class name.
Column J — order_name: Order name.
Column K — family_name: Family name.
Column L — iucn_cat: IUCN Red List category (LC, NT, VU, EN, CR, DD, NE).
Column M — country: Country at which each photograph was taken, generalized. To limit disclosure of sensitive species-occurrence information and identification of individual entries, this field is set to Not disclosed for entries involving IUCN-threatened species (VU/EN/CR, i.e. threatened = YES) and for contest winners; for these entries, geography is available only at the coarser Continent/Realm/Biome level (Columns F, S, T). Sub-national entries in the source data were normalized to country level.
Column N — group: Taxonomic group.
Column O — people: Whether the photograph corresponds to the People's Choice category or the Jury's Choice category.
Column P — winners: Whether the photograph was a winner or not (Yes/No).
Column Q — domain: Domain to which the species belongs (TERRESTRIAL/AQUATIC).
Column R — threatened: Whether the species is threatened or not (YES/NO/UNKNOWN).
Column S — realm: Realm category at which the photograph was taken, following Dinerstein et al. (2017).
Column T — biome: Biome category at which the photograph was taken, following Dinerstein et al. (2017).

File: `category_grouping_lookup.csv`

A two-column lookup table documenting how each original contest category was mapped to the generalized category_group used in Column E. Columns: original_category, category_group. Provided for transparency; it contains no row-level or individual-level information.

Data de-identification and sensitive-data statement

To reduce the risk of identifying individual award-winning photographs (and their photographers) and of disclosing sensitive species-occurrence information, the following generalization and suppression steps were applied to this version of the dataset. These build on earlier revisions in which the prize variable was removed and contest years were binned into 3-year periods.

Precise locality removed. The original free-text Location field (named national parks, reserves, islands, and cities; ~1,015 distinct values) was deleted in full. It was the primary vector for both species-occurrence disclosure and photograph identification, and it is not required for the biogeographic analysis, which operates at the Continent / Realm / Biome level.
Contest/award specificity reduced. The original 55 fine-grained contest categories — which included individually named awards and special awards (e.g., named portfolio, lifetime-achievement, endangered-species, and grand-title awards, some carrying an explicit year) — were collapsed into 12 broad, subject-based groups (category_group). Award-specific labels and embedded years were not retained in the released data.
Fine geography suppressed for sensitive and high-profile entries. country was set to Not disclosed for all entries involving IUCN-threatened species (VU/EN/CR) and for all contest winners. For these entries, geographic information is available only at the coarser Continent/Realm/Biome resolution. Sub-national entries elsewhere in the field were normalized to the country level.
No direct identifiers. The dataset contains no photographer names, photograph identifiers/URLs, exact capture dates, or prize/ranking details.
Data-quality corrections for public release. Character-encoding artifacts introduced during data extraction were repaired in the retained text fields, and an obvious value typo in the domain field was corrected.

Residual disclosure and analytical trade-off. Because the scientific object of this study is the species represented in the contest, species identity (sci_name / com_name) is retained at the species level and cannot be generalized without defeating the purpose of the dataset. As a consequence, some records remain distinguishable when species identity is combined with retained contest metadata. This residual is inherent to species-level representation data; it has been minimized by removing precise localities, generalizing contest/award categories, binning years, and suppressing fine geography for threatened taxa and winners, while retaining no direct personal identifiers.

References

Dinerstein, E., Olson, D., Joshi, A., et al. (2017). An ecoregion-based approach to protecting half the terrestrial realm. BioScience, 67(6), 534–545.