Search

This work presents a systemic top-down visualization of Bitcoin transaction activity to explore dynamically generated patterns of algorithmic behavior. Bitcoin dominates the cryptocurrency markets and presents researchers with a rich source of real-time transactional data. The pseudonymous yet public nature of the data presents opportunities for the discovery of human and algorithmic behavioral patterns of interest to many parties such as financial regulators, protocol designers, and security analysts. However, retaining visual fidelity to the underlying data to retain a fuller understanding of activity within the network remains challenging, particularly in real time. We expose an effective force-directed graph visualization employed in our large-scale data observation facility to accelerate this data exploration and derive useful insight among domain experts and the general public alike. The high-fidelity visualizations demonstrated in this article allowed for collaborative discovery of unexpected high frequency transaction patterns, including automated laundering operations, and the evolution of multiple distinct algorithmic denial of service attacks on the Bitcoin network.

Poor planning on the construction site at the ‘look-ahead’ planning stage (where information from diverse sources is integrated as plans are being developed for the next six weeks) often result on cost overruns and schedule delays. This thesis addresses the inefficiencies emerging from manual look-ahead planning by means of an information modelling approach to codify and validate detailed construction process information. The key contribution is a novel Linked-Data based Constraint Checking (LDCC) method to identify construction constraint violations from information distributed over multiple heterogeneous sources. LDCC can be integrated with machine learning methods to generate a constraint-free Look-Ahead Schedule (LAS) automatically. This approach can augment decision-making in look-ahead planning through data-driven constraint identification and automated LAS generation. This research comprises of three main studies. First, digital information use in three construction projects was studied to inform the development of the information modelling approach. This study identified three codification challenges—software usage, information sharing, and missing construction process information. Second, building on the understanding of the codification challenges, the LDCC method was developed to codify and validate detailed construction process information, including complex construction constraints distributed over multiple databases. Third, the LDCC method was successfully integrated with two machine learning methods (Genetic Algorithm and Reinforcement Learning) to automate LAS generation. When tested on a real construction project, the LDCC method identified all the constraint violations in the manually generated LAS. Also, LAS generation methods automatically generated conflict-free LASs significantly faster than manual methods. Both results demonstrate the applicability of the developed methods on real construction projects. In summary, the thesis extends existing knowledge in the construction informatics domain by demonstrating the benefits of using linked-data based methods to address the issue of data silos in construction information, and enabling the application of data-driven-decision support tools to aid look-ahead planning. Open Access