Last week, Coinalytics presented its technology, Blockstem, at San Francisco FinDEVr Conference.
Coinalytics is a venture-backed startup based in Palo Alto pursuing the mission of “providing real-time intelligence for decentralized blockchain platforms”.
Coinalytics’ unique approach applies advanced machine learning and state-of-the-art distributed systems to cryptographic platforms, enabling enterprises to derive real-time business intelligence and risk assessment from blockchains and decentralized applications.
This enables companies in a variety of industries like online consumer payments, financial services and decentralized applications to make smart decisions in real-time, while maintaining data privacy and trust.
Real-time intelligence leading to actionable insights significantly lowers the barriers for mainstream enterprise adoption of blockchain technologies.
Right at the intersection of the blockchain, Big Data and artificial intelligence, the platform employs advanced pattern recognition methodologies and real-time online learning to mine pseudonymous data with sparse features.
During the DEVr Conference, Coinalytics co-founder, Bill Gleim, explained there are fundamental differences between traditional finance and the bitcoin blockchain technology.
The financial industry has imperfect visibility of transactions and good visibility of clients, whereas the blockchain provides just the opposite.
Coinalytics intends to fill the gap with the Blockstem API which provides coverage tracking the bitcoin network and its transactions.
Clients can run SQL-like queries against Blockstem to access blockchain data.
For example, we may query to discover which blockchain transaction had the most outputs and who initiated that transaction.
In the demo of this process, Blockstem returned a single transaction which had 5,352 outputs. Bill and James proceeded with further queries and were able to find all of the bitcoin addresses controlled by the same entity that initiated that transaction by querying their cluster addresses table.
They then traced the most recent transaction from the cluster and revealed that it has the label F2Pool, which most likely is a mining pool sending out payments to its miners.
Essentially, Coinalytics has expanded the blockchain data model by providing additional aggregate statistics which add value for users by allowing them to build more advanced applications on top of the blockchain.
The Coinalytics system is tolerant to multiple zone outages, opens up blockchain engineering to the public, and allows developers to use Coinalytics databases to hold and index blockchain data instead of managing their own.
Bill and James also referenced their Anomoly Detection module for forensics. They host a live transaction stream and can identify any transactions in the wire which are anomalous- an important first step in AML compliance or forensic analytics.
To demonstrate Coinalytics Forensic Analytics tools, the team demonstrated their GUI which can track the flow of funds throughout the blockchain.
They gave the example of Ross Ulbricht making payments into various accounts belonging to DEA agent Carl Force, who was recently charged with extortion.
The Coinalytics shows 4 transactions in a GUI, all of which were made from Ross’s wallet but seem to be directed to independent accounts.
By zooming in or out on the transactions, a user can ultimately see that they are, in fact, linked in a single cluster and delivered to one common, identified individual.
This forensic analysis can be accomplished by 1 person in a couple hours with this tool, as opposed the several months that it took a government team.
Coinalytics has created and scaled a datacenter on top of bitcoin data from the blockchain, replete with a queryable interface and API for creating new apps.
Their work enables a wave of future technology which could affect everything from banking to law enforcement.