First, it is crucial to define the data mining term, which is defined by Geoffrey Holmes and Sally Cunningham as “a technique that explores a rich, relatively unstructured set of data and retrieves from it unusual patterns, unexpected regularities, implicit information, etc.” [3] This definition describes exactly why it can be so powerful to construction companies. It transforms the simple data into decision-making supportive information by detecting unusual aspects that will be useful in the identification of what is working and what is not. It is much easier to become a more efficient company when you really know what is going on.
Author: Joana Coutinho; António Ruivo Meireles
Country: Portugal
Company: ndBIM Virtual Building
- INTRODUCTION
First, it is crucial to define the data mining term, which is defined by Geoffrey Holmes and Sally Cunningham as “a technique that explores a rich, relatively unstructured set of data and retrieves from it unusual patterns, unexpected regularities, implicit information, etc.” [3] This definition describes exactly why it can be so powerful to construction companies. It transforms the simple data into decision-making supportive information by detecting unusual aspects that will be useful in the identification of what is working and what is not. It is much easier to become a more efficient company when you really know what is going on.
Everyday construction firms generate a big amount of data during their projects, which they store in their data systems throughout time, such as labor, machinery, materials, among others, related with the design, estimate, schedule, quality, etc. [1] This paper introduces two fundamental queries: the importance and the transparency of BIM information and the fact that the construction companies do not reuse any data to support their decision-making process.
It is increasingly apparent that Building Information Modelling (BIM) offers a vast amount of benefits to the construction industry, namely clash detection, better coordination, 4D scheduling, among many others but, in fact, the major benefit of BIM is the powerful information that is created along the process. Many companies are already adopting this technology but still do not reuse the information that results from it: productivities, task durations, costs, subcontractor information, and so on.
This methodology arises in a way that enables construction companies to start using all the stored construction data in order to support and enhance the decision-making tasks. As the project develops and changes are being made, all the information is updated and coordinated at real time while the relationships between the several information resources are expressed as a whole. [2]
The amount of information which results from the process, not only from the 3D design and coordination phase but also from the construction management stage when the project is planned and then monitored by the BIM platforms, must be manipulated to allow managers to see the kind of information they need at each moment. If the collected information is not standardized then it becomes useless and it is not possible to make any statistic study with it, remaining unsafe to use in future projects. Thus, it is important before any usage to categorize it by cost, task, subcontractor, etc.
It is also proposed a method which allies not only the benefits of BIM but also the data mining tools that could make the information valuable for managers to make better decisions.
- MATERIALS AND METHOD
The first step for reusing the construction information is to define a standard procedure for nomenclature and organization of the data that is extracted from the budgeting, scheduling and monitoring processes. The main procedure is represented in the following scheme.
As shown in Figure 1, managers can extract real information from management BIM software not only during (actual information) but also at the end of each project (historical information), such as labor productivities and consumptions, task durations, quantities, costs, and so on. There are two parallel paths in this procedure: actual and historical. In the first case, the information is extracted regularly during the project and then analyzed and presented on a dashboard. On the other, it is extracted from the several systems and then standardized, analyzed and stored.
The dashboards, both actual and historical, are simply a way of representing the construction and BIM data, as it is visible on Figure 2. This panels allows managers, technicians and CEOs to realize if the project is actually going according to plan (and its causes, if not) and then support their decisions on the facts that are presented on it.
What makes this procedure so unique and effective is the standardization of data presented on it. This is because the standardization is the keystone towards data manipulation, which must be correctly set in the procedure due to the fact that nomenclature, frequency of extraction and storing destination enable the data mining tools to filter and operate the information. There are infinite standards for information organization and structure and, in the BIM area, Omniclass and COBie, while different, are two of the most used. What this method argues is that you most standardize your data but it does not matter how you do it, you can do it with any standard you like that suits your projects or you may even create one for yourself.
There are many solutions for dashboards and the best way to evaluate their features is to develop a benchmarking study and score the several functionalities. The key requirements the platform must meet are: Excel import, preferably in the cloud for automatic updates; basic analysis functions; statistical analysis; and graphical and regression analysis.
In addition, there are several types of dashboards. Although the most common is the web-based application, some of them are an application that remains physically on the equipment (PC, tablet and/or smartphone) and the dashboard may even be developed using Excel, but without the data interactivity animations (drill down, drop down, filtering).
- RESULTS AND DISCUSSION
The information and values that outcome from the construction data mining are real and justified, because they are obtained from real productivities, durations, consumptions and many other items from past projects. In this way, it is much easier to identify problems (e.g., the task responsible for the project delay, the subcontractor that could not follow as predicted, etc.). Another big benefit of reusing construction data is to enable the creation of more accurate schedules and budgets, as the historical information is being updated at each project (actual information) and becoming more and more precise.
- CONCLUSIONS
It is concluded that there is a substantial waste of information regarding the construction industry, which turns more evident by an accurate BIM implementation. The reuse of the information, if properly done, can lead to a greater efficiency in the management of information and resources of a project, in order to support the decision making process. The use of a dashboard enables the representation of this data, automatically updated, in an extremely visual way and also it can be shared with all the departments. However, it is important to follow a strict procedure of data reuse and to actually take advantage of the information generated, while the representation platform is just one of the means needed to achieve a higher end.
One of the greatest difficulties is to ensure that the process is followed by the participants and that the standardization is performed correctly, which allows companies to take real advantage of data mining processes. In order to gain reliable values, the data must be extracted at the same time and in the same manner, in every site. Thus, the key challenge is to assure that there is discipline in the process, setting strict procedures and assuring that these are followed.
- REFERENCES
[1] Rujirayanyong, T.; Shi, J. A project-oriented data warehouse for construction. Automation in Construction, 16/11/2005, 800-807.
[2] Zhiliang, M. A BIM-Based Approach to Reusing Construction Firm’s Management Information. Australasian Journal of Construction Economics and Building, 2012, 29-38.
[3] Cunningham, S.; Holmes, G. Using Data Mining to Support the Construction and Maintenance of Expert Systems. First New Zealand International Two-Stream Conference, 24/11/1993, 156-159.