Home

Welcome to the Andromeda Docs!

This wiki serves as the documentation (user and admin guide) for Andromeda and related packages contained within this repo.

Overview of Andromeda

Andromeda uses a weighted multi-dimensional scaling algorithm to project high dimensional data to a 2-dimensional plot. This visualization is interactive, allowing the user to choose image representations as their points and rearrange the clusters, and dimensional weights, in a manner they find appropriate. This package has an additional 'Fetch Data' option which pulls from iNaturalist, and can be merged with additional satellite data. This includes integration with latloncover to classify the percentage of various categories of landcover within two scales centered on each observation geocoordinate: small (1/2 x 1/2 mile) and big (2 x 2 mile). See LatLonCover subcategories for the definitions of the classifications.

An Andromeda instance allowing users to perform dimensional reduction on an uploaded CSV file can be found on the Imageomics Hugging Face.

Andromeda Guides

Please see Analysis with Andromeda for more information on how to get started working with this human-in-the-loop visualization tool. The Fetch image data functionality is described in Fetching Data from iNaturalist. Information for developers can be found here and in the Developing with Docker section of this wiki.