Clustering and analysis of user behaviour in blockchain: A case study of Planet IX

Dorottya Zelenyanszki, Zhé Hóu, Kamanashis Biswas, Vallipuram Muthukkumarasamy · 2025

Read Clustering and analysis of user behaviour in blockchain: A case study of Planet IX by Dorottya Zelenyanszki, Zhé Hóu, Kamanashis Biswas, Vallipuram Muthukkumarasamy (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-11702.txt.

Introduction

Dorottya Zelenyanszki and three co-authors show how much player behavior can be mined from a blockchain game, in a 2025 paper. Every move in the game Planet IX is written to a public chain, open for anyone to trace. The team built a pipeline that harvests those transactions, rebuilds each player's sequence of actions, and notes how game tokens move through the flows. A graph neural network then embeds each player's flow so clustering algorithms can group similar styles.

Benchmarking several clustering methods, they found clear behavioral clusters with distinct habits. Visualized as graphs, the clusters reveal information an attacker could exploit, and the team sketch a threat model covering several harmed areas. The study is a warning: transparency in decentralized apps cuts both ways. What makes chains trustworthy also makes their users legible.

Catalog record

CREATORDorottya Zelenyanszki, Zhé Hóu, Kamanashis Biswas, Vallipuram Muthukkumarasamy
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE58 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-11702

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Clustering and analysis of user behaviour in blockchain: A case study of Planet IX Dorottya Zelenyanszki, Zhé Hóu, Kamanashis Biswas, Vallipuram Muthukkumarasamy Abstract Decentralised applications (dApps) that run on public blockchains have the benefit of trustworthiness and transparency as every activity that happens on the blockchain can be publicly traced through the transaction data. However, this introduces a potential privacy problem as this data can be tracked and analysed, which can reveal user-behaviour information. A user behaviour analysis pipeline was proposed to present how this type of information can be extracted and analysed to identify separate behavioural clusters that can describe how users behave in the game. The pipeline starts with the collection of transaction data, involving smart contracts, that is collected from a blockchain-based game called Planet IX. Both the raw transaction information and the transaction events are considered in the data collection. From this data, separate game actions can be formed and those are leveraged to present how and when the users conducted their in-game activities in the form of user flows. An extended version of these user flows also presents how the Non-Fungible Tokens (NFTs) are being leveraged in the user actions. The latter is given as input for a Graph Neural Network (GNN) model to provide graph embeddings for these flows which then can be leveraged by clustering algorithms to cluster user behaviours into separate behavioural clusters. We benchmark and compare well-known clustering algorithms as a part of the proposed method. The user behaviour clusters were analysed and visualised in a graph format. It was found that behavioural information can be extracted regarding the users that belong to these clusters. Such information can be exploited by malicious users to their advantage. To demonstrate this, a privacy threat model was also presented based on the results that correspond to multiple potentially affected areas. [orcid=0009-0004-1300-0357] [orcid=0000-0001-7164-0580] [orcid=0000-0003-3719-8607] [orcid=0000-0002-6787-6379] keywords blockchain ,Non-fungible tokens ,user behaviour ,clustering 1 Introduction Public blockchains have the benefits of immutability, decentralisation and anonymity [ 1 ] , and have a wide range of potential use cases, such as distributed identity management [ 2 ] , record linkage [ 3 ] and metaverse [ 4 ] . DApps that run on top of blockchains and NFTs that represent various types of physical and digital objects, such as arts [ 5 ] , have become widely used in recent years. However, every submitted blockchain transaction can be publicly traced by anyone, thus, the transaction data can be analysed, which can uncover behavioural information. This is somewhat analogous to being able to analyse people’s browser history and activities, which is a serious privacy concern. Zhang et al. [ 6 ] described this as the issue of linkability of transactions, which makes de-anonymisation inference attacks possible. There have been multiple research works conducted on blockchain transaction analysis, however, these have differing focuses and they usually do not involve the events that are emitted by the transactions in their analysis. Analysis in regards to NFTs is mainly limited to their market analysis and related works do not consider their actual usage in dApps. For instance, Tao et al. [ 7 ] presented a network analysis of the Bitcoin blockchain, and Pelechrinis et al. [ 8 ] provided an analysis of the NBA’s TopShot NFT marketplace regarding illegal activities such as money laundering or trading illicit goods. However, the type of user behaviour that can be extracted from dApps through transaction analysis remains unknown which prompted this research as there is a need to present the full extent of the previously mentioned privacy concern. Furthermore, behaviour analysis can be leveraged to identify malicious entities such as phishing scammer participants [ 9 ] . This work aims to analyse behavioural information regarding in-app activities, including NFT usage by users of blockchain-based games. We argue that this type of information can be considered problematic, as malicious users can use it to their advantage, for example, for impersonation attacks or targeted scams [ 10 , 11 ] . It can also be used by related parties such as game providers, to track the users and their activities and implement changes based on the corresponding analysis. Tracking users in blockchain networks can be done in multiple application areas. Hu et al. [ 12 ] studied tracing users for blockchain data audit purposes. In our work, a blockchain-based game called Planet IX 1 1 1 https://planetix.com/ was chosen as a case study application. Transactions from multiple smart contracts were collected and used to get the general transaction information and the emitted event logs. The event logs were decoded to obtain all the event data which was then converted into a graph format, thereby allowing it to be queried to access information easily. All related event sequences of users and NFTs were extracted through database queries, which allowed the formation of general in-game actions. From the event sequences and formed actions, the corresponding user action flows for every user were presented. Similarly, for every tokenId (which is a property that corresponds to an NFT), the associated actions were extracted. They were also linked to user wallet addresses, which enabled the presentation of extended user flows that not only show the conducted activities but also which tokens were used at what time by the performed actions. These extended user flows served as input for a GNN model that assigned unique graph embeddings for each of them. A graph embedding is a vector representation of the graph that is suitable to be used for clustering. Following that, several clustering algorithms were utilised to cluster the user flows into separate unique behavioural clusters. The algorithms were evaluated based on common metrics, and one of them was selected as the current method for the clustering process. The output behavioural clusters were then described which presented the separate types of user behaviours that are possible in this chosen blockchain-based game. A corresponding privacy threat model was also included to present how this can potentially be utilised maliciously. This research presents a new type of blockchain analysis that focuses on behavioural information that can inform users’ activities, interactions and habits. The malicious actors or dApp developers do not need to fully identify the person behind the user as they can construct a profile based on the behavioural information and use that to their advantage to conduct activities like targeting the original user based on past actions or using the constructed profile to perform impersonation-related attacks. Behavioural information does not directly affect users’ privacy but by linking multiple activities, relations and habits of users together, it can reveal sensitive information. Therefore, behavioural information is privacy-sensitive and has to be protected. The overall contributions of this research are summarised as follows: • Proposed an action synthesis method that presented unique actions from a blockchain-based application which were used to construct users’ action flows. • A user behaviour analysis pipeline was proposed that adopts GNN. Multiple clustering algorithms were utilised within the pipeline to cluster the user flows into unique behavioural clusters. These clusters can be used to analyse different user groups. • The results were compared from the clustering algorithms in order to identify the well performing ones for this task. Descriptions were provided

Read the complete text