Well it would be a package by itself which will be an [implementation] (What is Python Mapper? — Python Mapper documentation) of the Mapper algorithm. There exists an implementation but it is in Python so I guess that my main task will be to understand it and port it to Julia.
Skill Set :-
- Familiarity with writing code in Julia
- Familiarity with Topology and specifically Homology
- In-depth understanding of algorithms related to Persistent Homology and the Mapper algorithm
- Familiarity with graphing packages like Plots.jl
Frankly speaking I have never seen a project involving Topology in GSoC. It is a rather new and unknown domain which is currently dominated by a few experts, some dedicated research labs and a private firm.
That being said I’m attaching a sample of output generated by this algorithm.
This is a study on Asthma where the dataset is high dimensional. The nodes are not individual data points but they are small microclusters. TDA disregards the distance between individual points and aims to reveal the shape in the data which is independent of dimensions. Note that this is a graph and not a plot of the dataset.
