author:
(1) Prakhiyat Khati, Computer Science, University of Saskatchehua, Saskatoon, Canada ([email protected]).
Links table
One summary and introduction
2 related works
3 Datasets and experiment setup
4 Methodology
4.1 NFTs Transaction Network
4.2 Predicting the NFTs Bubble
5 Discussion, conclusions and references
4.1 NFTs Transaction Network
We created the interaction network to perform transaction analysis on the NFT network, based on the table of tokens extracted from Dataset 1 observed from the large query. This table is created based on token transfers from one address to another. Their address can be either the buyer or the seller. Some addresses 0x000 indicate mint and copy addresses. When an NFT is removed from the blockchain ledger, the destination address will be the copy address, and whenever an NFT is minted or created for the pool, its address will initially be 0x000 before the owner is assigned. Like our network, we build the transaction graph model into a weighted multidirectional graph MDG(V, E) with a set of nodes V and edges E. Each node represents one wallet address; It can be considered a user address. Each directed edge E represents one token transaction between nodes V. For each transaction, information is specified and stored as edge parameters. For example, transaction cost_hash_id paid by the wallet, Date_time of the transaction. Here we used the Python package “igraph” to extract relevant information from the spreadsheet.
NFT sales skyrocketed in 2021; Figure 6 shows the average sales quantity over time. We can note Table 1, which shows the increase in the number of transactions during the year. In 2018, there were only 450,000 transactions, while in 2021, the number of transactions reached nearly three and a half million. If we look at Table 1, we can see a boom in the NFT world as only in the first quarter; The total volume has risen to approximately US$760 million. Here the in-degree and out-degree of each node are too large to be visualized; Figure 7: Shows the top 10 high value accounts. This high value means that these addresses are actively trading NFTs. The highest number of NFTs, as shown in Figure 7, are located under an address starting at 0x0000. This means that any user does not own this address; These titles are associated with NFTs when they are first minted.
We have selected the top 7 NFT projects [37] at all times according to the number of sales during the period and calculate the semantics of the combined network by performing histogram analysis on the transaction data. We observed that there are a total of 33,678 unique addresses, which represent nodes of the graph, and similarly 425,246 number of transactions between those addresses.
The constructed network is a directed graph. We compute the reciprocity of the graph network, the interconnectivity, connected components, and k-core properties of the network. The reciprocity of a directed graph obtains the probability that nodes in the directed network are mutually connected.
The ratio here is the number of edges pointing in both directions to the total number of edges in the graph. This shows the business transaction between user nodes. Likewise, in the above equation. Reciprocity is defined for one node u. It is the ratio of the number of edges in both directions to the total number of edges connected to the node. G denotes a directed graph. It was noted that the reciprocity in the graph was 0.068361305. Likewise, among the nodes, we found that 18,279 strong nodes and 72,922 strong edges were connected between the nodes, and we observed that the maximum number of weak nodes was only 95 nodes; This shows how well all nodes are connected. We note that the maximum number of edges of weakly connected components was 104064. Table 2 gives all other graph parameters observed from the histogram of NFT transactions. Table 2 shows the number of strong and connected nodes and edges. Diversity is calculated in a non-trend graph.
Here we first combined the seven different CSV files into one group and assigned a unique ID to each title. The giant() method was used to obtain the strongly connected components, which is similar to the weakly connected components. Then we calculated the core of the large weakly connected components and used the k_core method to get the number of vertices of the strongly and weakly connected components and the number of edges.
The figure below shows the entire NFT network component as shown here, where the buyer and seller are generated from the merchant, and the transaction is considered a component for better visualization.
As shown in Figure 2. Let us take an example to visualize the transaction of a single NFT transaction during this period. As we see in the figure below. The pink is the NFT called “See You Later,” the red is the NFT created and then all the orange nodes are the buyer and seller from the time the NFT was created. The blue color represents the transaction and the type of transaction that occurred. For example, it shows which wallet sold an NFT and which wallet bought an NFT from that transaction.
We can see that the graph contains the components of the group. Most common NFTs belong to a collection. We can see the example in the real world where most famous paintings by a famous artist are sold at high prices. In a similar way, the NFT collection and who created the NFT determine the value of the NFT collection. There is no price limit as there is no requirement on how much price can be set.
Likewise, in order for InSite to also influence nodes in the NFT network, we used the Article Ranking algorithm, which is a form of the PageRank algorithm. Measures the transitive effect of a contract.
Neo4j provides us with built-in methods to calculate the rating. Here article rank reduces the influence of low-degree nodes by lowering the score sent to its neighbor in each interaction.
Tables 5 and 6 show the most influential NFT collectors in the network and their trading value. The volume of sales and purchases in US dollars. This indicates that there is a significant amount of transaction influence through these nodes. In Figure 9 we can see a community of these influence nodes distinguished using a different color. We can observe a denser community in the center and many high influence nodes and multiple small communities being formed around it, these nodes are densely connected which indicates high transactions between nodes.
