I wanted to share some results on a recent paper I submitted to arxiv (will be available early next week). I am proud to say that 100% of the data is sourced entirely on-chain from Subgraphs.
The research paper is primarily the result from the ability to build a data pipeline to obtain on-chain data from the Subgrounds library. As a result, I am able to integrate it into a wider data analysis pipeline and prototype before moving the model to production.
I used the Mapper algorithm, a topological data analysis (TDA) tool, to identify six statistically distinct MEV behaviors within an Olympus MEV dataset, curated by Subgrounds.
The specific model is Mapper algorithm, which is a topological data analytics tool. It transform scontinuous spaces of datasets to discrete graph structures while preserving topological and geometric information. I wrote a preview post for the arxiv paper on mirror which contains charts and a paper overview for the interested reader.