Read the success stories of crypto entrepreneurs.

Discussions are fun when we are part of a community.
Login Free Registration

Get 10 AltcoinN Points just by registering on this forums.

Centralized Crypto Exchanges Explained in 5 Fascinating Data Visualizations

Centralized exchanges are, arguably, one of the most challenging components to analyze in the crypto-asset markets. While the behavior of many actors in the crypto space is transparently recorded in distributed ledgers, centralized exchanges still operate largely off-chain only publishing subset of the activity to the corresponding blockchains. Without a doubt, centralized exchanges introduce a level of opacity that challenges even the most sophisticated analytic techniques. And yet, the analysis of the behavior of centralized exchanges can yield many interesting benefits for crypto investors and traders. Imagine that you are able to effectively track large crypto transfers between exchanges that can anticipate a large position on a specific crypto asset. All that, however, requires understanding the underlying patterns in centralized exchanges.

At IntoTheBlock, we have been actively working on really advanced machine learning models that help us understand the behavior of known actors in the crypto space including centralized exchanges. Despite the intelligence of the machine learning models, many times we need to rely on human-centric data exploration to validate some of the results. To address that challenge, IntoTheBlock’s technical lead extraordinaire Pablo Biancotto led the implementation of an internal graph explorer tool that help visualize relationships between addresses, transactions and other known entities in the blockchain ecosystem. Some of the visual exploration reveal fascinating patterns that take place every day in centralized exchanges.
A Quick Intro to Centralized Exchange Topologies
Understanding the behavior of centralized crypto exchanges requires analyzing some of their key components individually and as a group. At a high level and generalizing a few concepts across exchanges, there are a few key components that are relevant in the behavior of centralized crypto exchanges:
· Hot Wallets: Hot wallets are typically the main interaction point between external parties and an exchange. Exchanges use this type of wallets to make an asset available to trade.
· Cold Wallets: Exchanges use cold wallets as a secured storage of crypto-assets. This type of wallets typically hold larger amounts of assets that are not intended to be traded frequently.
· Deposit Addresses: Deposit addresses are, often temporary, on-chain addresses used to transfer funds into an exchange. The focus of this type of address is to facilitate user to exchange money flows.
· Withdrawal Addresses: Withdrawal addresses are, often temporary, on-chain addresses that are used to transfer funds out of the main exchange wallet. Sometimes withdrawal addresses can play a dual role as deposit addresses.
While those four components represent the core of the on-chain architecture of centralized exchanges, identifying them requires fairly sophisticated heuristics or machine learning methods. Part of the challenge is that the patterns of interactions between these components can be arbitrarily complex and are still not very well understood. With the help of one of IntoTheBlock’s exchange machine learning classifiers, a visual exploration of the blockchain datasets reveal some fascinating patterns.
Visualizing Crypto Exchanges in the Blockchain
To understand the behavior of the key components of centralized crypto exchanges, let’s explore some fascinating visualizations that highlight some of their key patterns of interactions.
1)Deposit Addresses Transferring Funds Into the Exchange Main Wallet
One of the main patterns in crypto exchanges is the transfer of funds from deposit addresses into the exchange main wallet. The following visual clearly illustrates that pattern with the blue circles representing deposit addresses and the green circle representing the main wallet. A curious aspect to notice is how the funds of many deposit address get bundled in a single transaction between being transferred to the main wallet.
[Image: 40ymA95p0lcRGbZcoWg7Dwyzpw62-z72358f]
The same pattern can be observed at more scale in the following visual:
[Image: 40ymA95p0lcRGbZcoWg7Dwyzpw62-jd1w35if]
2)Exchange Main Wallet Distributing Funds to Withdrawal Addresses
Another traditional pattern in centralized exchanges is to send transactions to withdrawal addresses. The following visual illustrates that pattern with the orange circles representing the withdrawal addresses and the green circle the main exchange wallet.
[Image: 40ymA95p0lcRGbZcoWg7Dwyzpw62-0h3435wh]
3)Withdrawal Addresses that Also Act as Deposit Addresses
In some scenarios, withdrawal addresses also act deposit addresses, very often from other exchanges. That pattern is illustrated in the following visual. Blue circles represent deposit addresses and green circles represent withdrawal addresses.
[Image: 40ymA95p0lcRGbZcoWg7Dwyzpw62-pa4b3511]
4)Inter-Exchange Transfer
The same pattern illustrated in an inter-exchange transfer between Poloniex and Binance.
[Image: 40ymA95p0lcRGbZcoWg7Dwyzpw62-3n5i35o2]
5)Co-Spent Transactions for Withdrawal
Unspent transaction output (UTXO) are an important pattern in cryptocurrencies like Bitcoin, Litecoin or Bitcoin Cash. From the exchange standpoint, many transactions used unspent outputs as inputs to specific transaction. In the following figure, the little pink circle represents an unspent output that is combined with funds from the main exchange wallet in order to distribute funds to withdrawal addresses. Notice that another unspent output(big pink circle) gets distributed to a different address.
[Image: 40ymA95p0lcRGbZcoWg7Dwyzpw62-yn6p35yo]
As you can see, the combination of machine learning and advanced data visualizations reveals fascinating patterns in the behavior of centralized exchanges. Sometimes clever data visualizations spark the magic of human intuition to identify scenarios that are missed by the machine learning models. Centralized exchanges remain as one of the black boxes of the crypto-ecosystem but a little bit of machine learning and data visualizations will little by little help to unlock their secrets.