Imperial College London

Codification challenges for data-science in construction?


Construction sector taking up new forms of data

Codification challenges for data-science in construction?

Our research - Codification challenges for data-science in construction?

What is the research about?

What is stopping the construction sector taking up new forms of data science, including machine learning and data analytics? We argue that there are codification challenges. Our research examines information flow in three projects using Building Information Modelling (BIM) and a Common Data Environment (CDE) in the late design and construction phase to identify these challenges.

Why is it important?

Construction productivity can be significantly improved. Data science can help the teams that build our buildings and infrastructure to understand patterns in their behaviour, and to use this understanding to streamline and automate repetitive and dull processes, to focus more energy on value added activities. Yet, the lack of machine-readable information is limiting the application of data science.

What is the novelty?

The novelty is in identifying the significant challenges of codification that need to be overcome. These are software usage, information sharing and construction process information. Software usage challenges include interoperability, information loss during the conversion, and multiple modelling techniques. Information sharing challenges include unstructured information sharing, document control workarounds, lack of process change, and multiple common data environments. Construction process information challenges include the loss of constraints information and low level of detail. This also paper contributes by relating these to data quality dimensions, such as accuracy, completeness, accessibility, consistency, timeliness, and provenance.

What are the practical implications

In order to benefit from data science, construction practitioners should focus on data quality and the development of machine-readable information. While product information has become more structured through the use of BIM, three areas to consider are software usage, information sharing and the construction process information.

Who was involved

Ranjith Soman

Ranjith Soman
PhD Student - Research page


Professor Jennifer Whyte
Centre Director, Laing O’Rourke / Royal Academy of Engineering Chair in Systems Integration

Free access to the full paper (author copy)


Tim Gordon

Tim Gordon
Department of Civil and Environmental Engineering

Click to expand or contract

Contact details

Tel: +44 (0)20 7594 5031

Show all stories by this author