Global Nature

CODE: GitHub
DATATYPE: images
SOURCE: Instagram
PLACE: Bangkok, Berlin, Moscow, Sao Paulo, Tokyo
YEARS: 2014-2015
SIZE: 4,121
SUBSET: "nature"
VIZTYPE: growing entourage
FEATURES: stylistic; hierarchical clusters
FEAT SRC: Miriam Redi
PROJECT: Redi, Crockett, and Manovich. ACMMM '16
MORE INFO: Software Studies Blog

This is a "growing entourage" plot of Instagram images in the category "nature". Such plots take images located in a high-dimensional similarity space, cluster them by some method, and project the clusters to a 2D canvas. Here, the images are grouped by agglomerative hierarchical clustering in a 95-dimensional stylistic similarity space. The centroids of each cluster are then projected to 2D by PCA, and the centroids then take turns adding members to their "entourages", in order of visual similarity to the centroid. New members occupy the open square nearest their centroids. Territorial conflicts are settled by this principle. Empty squares are centroid locations. This plot has a tight grid, yielding a large central mass of clusters.