This repository contains the full dataset and analysis code accompanying the preprint:
"Artists Can Design Memory, and AI Can Predict It" (https://osf.io/preprints/psyarxiv/3g5az_v2)
This folder includes all data used in the analyses reported in the manuscript. It contains:
- Artwork-level metadata (e.g., artist background, contest category, size, medium)
- Low-level image features (e.g., hue, saturation, luminance, edge density)
- Spatial layout data from the public gallery exhibition
- Behavioral results from online (Exp. 1) and in-gallery (Exp. 2) memory experiments
- Subjective human ratings (Exp. 3) across 21 attributes (e.g., beauty, complexity, emotional valence, reconstructability)
- AI-based memorability scores from ResMem
The data file is stored in .csv format and is analysis-ready.
High-resolution versions of the accepted contest artworks are available via the project’s OSF repository (https://osf.io/y3mdr/files/osfstorage), under the folder Artworks.
- 86 artworks were used in the three experiments and analyses described in the manuscript.
- 4 additional artworks are included for completeness but were not used in any analysis (e.g., due to disqualification or artist withdrawal).
Note: Each image filename in the Artworks folder is indexed to match the Image_Num column in the data file (All_Predictors.csv), allowing direct correspondence between artwork images and their associated metadata and behavioral results. Because there is no image numbered 73, the artwork numbering extends to 91 even though there are only 90 artworks. The final four images (Image_Num 88–91) correspond to the four additional artworks that were excluded from all analyses.
This folder contains all Python notebooks used for:
- Running statistical analyses (e.g., regressions, clustering, PCA)
- Generating all figures included in the main text
Each notebook is thoroughly documented and organized to follow the structure of the manuscript.
If you use this dataset or codebase in your own work, please cite our accompanying manuscript.