This repository was created to support the article Food Data Analysis using Multidimensional Visualizations based on Point Placement written by Maria Eduarda M. de Holanda and Vinícius R. P. Borges. FAPDF and ProIC/UnB supported this research.
The selected dataset can be found in vegan_dataset.
Four state-of-the-art and recent visualization techniques were considered in the proposed method: Principal Component Analysis (PCA); t-Distributed Stochastic Neighbor Embedding (t-SNE), Uniform Manifold Approximation and Projection (UMAP) and TriMap. The implementation in Python can be found in sibgrapi_visualizations.ipynb.
In order to evaluate the quality of the selected visualization techniques, we followed two steps:
- finding the best choices of hyperparameters in each method, that can be found in sibgrapi_hyperparameters.ipynb.
- comparing the results with the previous mentioned evaluation metrics, which can be found in sibgrapi_evaluation.ipynb.