PEOPLE On-chain Transaction Network Analysis and Valuation
Analysis and Valuation of PEOPLE's On-chain Transactions
Original author: zelos
TL;DR: This article will start with PEOPLE, explore the history of PEOPLE’s transactions, analyze the behavior of traders and the characteristics of the transaction network. We also try to propose several factors to explain the price of Meme coins.
Meme coin usually refers to cryptocurrencies that have no practical use cases or application value. These currencies are mainly driven by the community and have high volatility compared to mainstream cryptocurrencies. When Meme coins gain community approval and trigger FOMO emotions, they may skyrocket in price overnight, bringing thousands of times of returns. A few investors get rich as a result, but many investors lose money due to market fluctuations and some scams.
Unlike most Meme coins (such as PEPE: Sad Frog and DOGE: Shiba Inu) generated by internet Meme culture, the creation of PEOPLE has specific use cases. ConstitutionDAO launched a crowdfunding campaign in November 2021, attempting to bid on a manuscript related to the US Constitution. Participants donated Ethereum and received PEOPLE tokens as decisions for future constitutional purposes. After the auction failed, all participants can exchange PEOPLE back to Ethereum at the original ratio of 1000000:1. Therefore, PEOPLE is deflationary, its quantity is continuously decreasing, and it has a theoretical minimum price. This makes it different from other Meme coins. This article will start with PEOPLE, explore the history of PEOPLE’s transactions, analyze the behavior of traders and the characteristics of the transaction network. We also try to propose several factors to explain the price of Meme coins.
Data Introduction
- Conversation with Electric Capital Partner Ken Deeter: Investment P...
- Variant Fund: Three potential directions for Web3 advertising
- Binance Research Interview with Cryptocurrency Institutional Invest...
The raw data comes from transaction information and log data on the Ethereum chain. We obtained a total of 157,597 transfer transactions involving PEOPLE tokens on Ethereum from November 15, 2021 to May 22, 2023 by crawling EVENT and parsing and converting it, recording the transaction timestamp, sender, receiver, and transaction amount. These transaction records involved a total of 46,847 transaction addresses. We cleaned and filtered out duplicate nodes on the transaction intermediate path and their corresponding transaction data.
Transaction Types
By analyzing the transaction data fields, we can classify each transaction to distinguish its behavior. Among them, transfer refers to token transfer transactions, which are the transfer of PEOPLE coins between user wallets; DEX_xxx refers to PEOPLE coin transactions conducted by users through decentralized exchanges, such as Uniswap, 1inch, DODO, etc.; mint and burn refer to users minting and destroying PEOPLE tokens; uni_burn and uni_mint are transactions in which users participate in providing liquidity; batch_transfer is a batch token transfer transaction that transfers tokens to multiple addresses at once; and MEVBot refers to transactions executed by robots or programs.
Based on the type of transaction, we can see that the transfer of PEOPLE tokens between user wallets accounts for the vast majority of transactions on the network, followed by PEOPLE transactions conducted by users on Uniswap.
Address Type
In order to explore the nature of PEOPLE transactions on the graph database, we integrated the addresses involved in the transactions as nodes in the graphFrame and distinguished whether the address is a contract address or a user address by checking if it has code. Contract addresses contain Solidity smart contract code; otherwise, the address is an EOA.
Address Yield
We summarize the transactions in which each address participates, determine the change in PEOPLE for that address, and calculate the ΔETH for each transaction. MINT 10 PEOPLE corresponds to a ΔETH of -10*1/1000000. For transfer of 10 PEOPLE, the ΔETH is calculated based on the PEOPLE-ETH price at the end of the block in which the transaction occurs. We accumulate ΔETH in transaction order, and the lowest value in the accumulated ETH sequence is the total cost. Next, we settle at the last transaction of the address, calculate the ETH and PEOPLE balances, and convert PEOPLE to ETH at the settlement price. The balance/cost is the yield for this address in PEOPLE transactions.
In this calculation method, we assume that users prefer to use the proceeds from selling PEOPLE to reinvest in PEOPLE. We did not consider the fee income from liquidity mining. The income calculation is based on ETH standard. In order to intuitively calculate the income situation, we did not convert the income to annualized income.
Analysis
Network Structure
In order to explore the network structure of PEOPLE token transactions, we studied the degree distribution of the PEOPLE network. By observing the degree frequency distribution of the undirected graph, we found that nodes with smaller degrees account for the vast majority of the degree frequency distribution. Therefore, we used a power-law distribution fit to explore the scale-free nature of the PEOPLE network (Figure 2). “Scale-free” means “lack of intrinsic scale,” which is the result of the coexistence of nodes with very different degrees in the network [Ref]. The degree distribution of the PEOPLE network can be approximated as: , where k is the degree of the node, and is the frequency of nodes with degree k.
After taking the logarithm, we can estimate the λ value of the undirected graph through linear fitting: 5.5563; the λ value of the user-to-user transaction network: 5.4054; the λ value of the user-to-contract interaction network: 4.8768. According to the classification of scale-free networks, the PEOPLE network does not have a significant scale-free property [Ref], and belongs to a small-world [Ref] network structure [Ref].
PEOPLE network has a highly aggregated characteristic, with many closely connected nodes. The average path length between user nodes is short due to a small number of hub nodes (e.g., Uniswap, empty addresses). In contrast to the scale-free condition, although hub nodes exist in the PEOPLE network, they are not large enough to significantly affect the average distance between nodes. In other words, most users will transfer PEOPLE coins with a specific address, and there is no particular trading address that has a significant impact on user trading choices.
We use closeness centrality to calculate the central tightness of the PEOPLE network, with an average centrality value of 0.21 (centrality=1 for connected undirected networks). Compared to other real networks, this indicates that most user addresses in the PEOPLE network will not only choose one address for trading, and the network is relatively robust.
The degree distribution scatter diagram of different categories of transactions, and the regression fit of the degree index λ, λ=5.5563
Frequency distribution diagram of closeness centrality, mean=0.21
We explored the dynamic changes of the network structure in time windows. Taking 216000 time blocks as a time window, which is roughly a month on Ethereum, we calculated the network composed of transactions in this month, and then took steps of 10000 to move the time window on the time axis to observe the changes in user outdegree and indegree. The following figure shows the trend of PEOPLE price and on-chain transactions, as well as the average degree (indegree + outdegree) and number of nodes in the PEOPLE transaction network.
Address characteristics and benefits
We first counted the life cycle and maximum holding of each address in the PEOPLE transaction network. The start block is the block number where the user first traded PEOPLE.
50% of addresses hold PEOPLE for no more than 2 days. 7% of users hold PEOPLE for more than 200 days, which includes exchange addresses and wallets. From the perspective of the first purchase time, the earlier the address holding PEOPLE, the longer the holding period. Obviously, for early buyers, PEOPLE is more of a souvenir than an investment. The holding time and maximum holding of EOA addresses are not significantly related. The holding period of contract addresses is positively correlated with the maximum holding.
Define the average transaction interval = EOA address life / total number of transactions. We call the top 25% of addresses Blocking per hand, and the largest 25% diamond hand. Diamond hand addresses have longer transaction intervals. We classified users based on activity, whether they provide liquidity in Uniswap, and whether they are early participants (Minters) in PEOPLE, and calculated the profitability of different users:
The results show that early participants have considerable profits. Even without considering the handling fee income of LP, the average income of addresses participating in LP is higher. Users participating in ConstitutionDAO, as Minters, have considerable consensus income: an average rate of return of 14.01%. Experienced users with longer transaction intervals also get higher returns.
Community division
We turned the PEOPLE trading network into a directed, multi-link graph, and based on the characteristics of Meme coins spreading intensively through social networks, we used label propagation algorithm to divide the community. First, a different label is randomly assigned to each node in the network, and then the neighboring nodes are assigned weights based on the current label and the direction and amount of transaction in and out of the adjacent node. Labels with larger weights will cover smaller ones. After continuous updating and iteration, nodes with the same label are divided into a community.
After 10 iterations of updating, we got 11,454 communities, and the community results did not converge, but two communities had a very large number of people, 22,665 and 15,727, respectively, and the average number of people in the other communities was about 1.2 people. We classify addresses outside the two large communities as the third category. In the process of visualizing using neo4j’s neovis.js, we merged the multiple links between two addresses into one and added the amount. The size of the nodes in the figure represents the Blockingge rank value, the color represents the community type, and the thickness of the lines represents the amount.
The central node of the first community (binance 14 address) and part of the addresses with a distance of 4 from the center
The center node of the second community, as well as the partial address with a relative distance of 4 from the center
We calculated the Blockinggerank values of the top ten central nodes in each community based on transaction frequency and amount. The first community is centered around cross-chain transaction addresses of several exchanges, such as Binance, Okex, and Gate.io; the second community is mainly centered around on-chain transactions, such as Uniswap, 0x, and 1inch; the third type is more scattered user addresses.
The center point of the first community stands out, with an interaction frequency of 13,130 times for the 14 Binance addresses. From the scatter plot of daily transactions (Figure 4), its addresses remain active for a long time and are not as volatile as overall transactions. Among the top three exchanges, gate.io has the fewest number of transactions and the largest single transaction amount. The average return rate is 2.2546, the Minter percentage is 9.1687%, and liquidity providers account for 0.3353%.
The second community is mainly based on on-chain DEX transactions, and the top few are trading robots active on DEXs such as Uniswap v3 and Uniswap v2, with the first address having a transaction frequency of more than 50,000 times. The average return rate is 9.0938, the Minter percentage is 64.8076%, and liquidity providers account for 2.9062%.
From a statistical perspective, the label propagation algorithm has relatively successfully divided the two communities of users who interact with wallet addresses related to exchanges and those who interact on-chain. Moreover, the minter content of the second community is much higher than that of the first community, and the yield is higher than that of the first community, which is consistent with our previous calculations and observations. Therefore, these two types are likely to have significant differences in network structure and observable different transaction patterns.
Scatter plot of daily average value of the top three important addresses in community 1
Percentage distribution of holding period in different communities
Valuation
The value of Meme coins mainly comes from their social attributes, speculative opportunities, and value storage. For Meme coins with other commercial operations, their value also has a commercial premium, such as DOGE.
value = social attribute + speculative opportunity (+commercial premium) (+value storage)
We introduce Zipf’s Law to measure the social attribute of PEOPLE, using the number of addresses holding PEOPLE coins as the number of active users. The more users holding coins, the stronger the consensus and the more valuable the trading network. Speculative opportunities are reflected in the changes in market FOMO sentiment. In this analysis, we use on-chain transactions and the PEOPLE-USDT trading pair on Binance to measure market FOMO sentiment.
valuation = N logN (1+emo)
emo = (daily transaction volume – average transaction volume over the past five days) / total circulation
The results show that this model can predict the general trend of market prices and capture the timing of price increases. The limitations of the model are that it mainly uses on-chain data. After PEOPLE is listed on an exchange, trading moves from the chain to the exchange, increasing derivatives. Therefore, the model does not obtain complete information. The model also does not measure the value of holding PEOPLE as a souvenir, that is, the impact of addresses that simply hold PEOPLE for a long time on the price. The model is highly dependent on PEOPLE trading volume. In the later stage, when PEOPLE trading is inactive and liquidity is low, it is difficult to estimate its value.
Summary
PEOPLE’s on-chain trading network does not have an obvious scale-free property, but it still conforms to the small-world network structure: highly clustered, but important central nodes (such as exchanges and pools) have a small impact on transfer users. In addition, the community division results show that there are behavioral differences in the network structure between users who frequently interact with exchange addresses for PEOPLE trading and those who trade PEOPLE only through on-chain addresses.
Users who participate in PEOPLE as Minters have received considerable average returns. Addresses that did not participate in Mint can only gain profits by early participation. Half of the addresses holding PEOPLE only hold it for two days. PEOPLE trading was only active in the early stages and has been in a downtrend ever since. Obviously, PEOPLE has passed the stage of active trading and high speculative value and has entered the later stage as a Meme coin. Most of the current holders regard PEOPLE as a souvenir and are not pursuing profits.
PEOPLE is a relatively pure M project, and its initial participants did not participate solely for speculation. The development cycle of the project is also complete. Based on PEOPLE, we will analyze other Meme projects and improve the valuation model in the future.
