Temporal analysis of the ether-flow network on ethereum

Abstract

Ethereum has grown significantly since its launch in 2015, driving more in-depth analysis into its network properties in the past few years. In particular, the vast amount of information available on cryptocurrencies has enabled researchers to analyse the network from a graph theory perspective. Majority of the studies are focused on the smart contract ecosystem while a few have delved into deanonymization techniques and link prediction. However, financial transaction networks, including cryptocurrencies, are constantly evolving and static graph measures are not sufficient to capture the dynamic properties of the network. This project analyses the temporal aspect of the Ether flow network from 2015 to 2020 using several metrics to characterizes user behaviour and network properties. This is followed by an in-depth analysis of Ether flow on Binance, a centralised cryptocurrency exchange, from 2017 to 2020. By incorporating the time aspect into network analysis, this project compares and contrasts user behaviour during different time periods of the Ethereum network. Ether transactions are collected from a well-maintained public dataset on Google’s BigQuery platform and the Python-based NetworkX library is used to analyse properties of the network. A key finding for this project is that the price of Ether might be a key factor for determining user behaviour on exchanges but is not reflective of overall user behaviour on the network.