Description of the Interactive Tool
The Embedding Projector is an open source tool for data visualization in high dimensions. There are example datasets based on words, images and iris dataset. Users can also upload custom datasets. The tool helps to find relations between samples and their neighbors (similar samples). For example, using word dataset, the tool displays the following words as similar to the word 'data': 'information', 'instructions', 'files', 'register', 'signals', etc. The projector also enables to compare different dimensionality reduction techniques (PCA, UMAP, T-SNE) and get insights on the machine learning process. The tool can be used both to analyze various types of data (numerical data, words, images, etc.) and to compare techniques.
More information about this tool is given in:
https://youtu.be/wvsE8jm1GzE<br />

Points of Strength
The tool encourages students to explore the example dataset, modify parameters of the dimensionality reduction methods and analyze the consequences. The large amount of exemplary data ensures ability to perform different experiments and perform analysis from different perspectives.