I'm very new to data science. I'm wondering how is it possible to graph data of n dimension. For example by Clustering high-dimensional data
https://en.wikipedia.org/wiki/Clustering_high-dimensional_data
How do we graph these data points? Do we use a specific data structure? Do we take only the three more relevant dimensions?
There are many techniques. They usually fall within the category of "manifold learning" and "dimensionality reduction" techniques. In both cases, it is assumed that the data lives in a much lower dimensional manifold as the original data.
For the purpose of visualization, you look for a mapping of your data to the plane, such that vicinity (topological distance) is preserved. That is, points mapped close to each other, are also close to each other in the original space.
One recent, very popular technique is tSNE, but there is a bunch of others.
Principal components analysis is always an option for reducing the number of variables you need to display. Alternatively, you can use things like point (on a graph) size, color, and shape to differentiate among different kinds of categorical and metric data. In R this can be done in ggplot.
As to clustering itself, I am unsure.
