This implementation uses NetworkX: https://github.com/networkx/networkx, Pandas, Maplotlib, SigmaJS and Python-Louvain for simple community arranging. Still a work in progress, but this can already produce interesting network analyses

Please donate some BTC to support the project: 1DbTky3KKfY5aShvHy4mrsSzHQZhsPr3zK

You can download the original thesis here

Due to the public nature of transaction data on blockchain based financial systems, it is possible to model these systems as a network and analyze its structure to provide and define whether the different variables that emerge in the transaction history of ERC20 tokens can be used to correlate the health of the system and therefore propose a risk coefficient that quantifies how active the token is, how distributed it is and infer how likely it is to respond to contagion and how stable the price is.

As is standard everything you need is in requirements.txt.

If it's your first time using vagrant, check out how to use it here.

vagrant up
get server running and start creating stuff
vagrant ssh

$ cd /vagrant/
$ python3.6 # To run a console

Instantiate SQL tables in a python console.

from schema.DBManager import DBManager
db = DBManager()
db.create_searches_table()

Searches are persisted by default in "searches" table that you instantiated. You can do other things if you need to, the db is called "mydb", to load psql console type:

psql -d mydb

Use of this database is optional and might not be used, however it might be practical to save large amounts of data and perform local queries to save costs on BigQuery.

Note that this software uses Google's BigQuery Ethereum dataset, based on the fantastic work found of Evgeny Medvedev: https://github.com/blockchain-etl/ethereum-etl.

You should activate a BigQuery account following these instructions: https://cloud.google.com/blog/products/data-analytics/ethereum-bigquery-public-dataset-smart-contract-analytics

and you should have your access info in ri5c/config/big_query_cred.json

PS: the location of that file is specified in an environment variable called GOOGLE_APPLICATION_CREDENTIALS

Find an interesting contract: (OmiseGo) omisego = "0xd26114cd6ee289accf82350c8d8487fedb8a0c07"

And now, from Python console run:

Use bundled minified SigmaJS data or go to their repo for an updated version, then run a simple Python server on your root folder:

python3.6 main.py

And go to 127.0.0.1:8000 to view the results

Choose a contract and type it into the "Get Contract" box.

from ri5c.get_contract import network_this

network_this(omisego, limit=3000)

Where limit is the transaction limit.

This will save the prepared data as network.png on your source folder

Note: These are examples graphed with Maplotlib, current iteration graphs using SigmaJS

omisego = "0xd26114cd6ee289accf82350c8d8487fedb8a0c07"

Data retrieved on March 2019:

• Transaction limite: 2,500
• Nodes: 2544
• Edges: 2489
• Average degree: 1.9568

cryptokitties = "0x06012c8cf97bead5deae237070f9587f8e7a266d"

Data retrieved on March 2019:

• Transaction limit: 3,000
• Nodes: 1625
• Edges: 1936
• Average degree: 2.3828

This is an experiment with weighted graph. tatatu = 0x9cda8a60dd5afa156c95bd974428d91a0812e054

Data retrieved on March 2019:

• Transaction limit: 2,500
• Nodes: 1776
• Edges: 1812
• Average degree: 2.0405

Feel free to collaborate, there are some issues and a project board. Please fork and grow. This code is open source, check the license