Data Quality: What is it and why is it important?

We often hear, even on a project with a theoretical BIM process, the comment: “Our contract is to deliver a 2D drawing output, so why do we need to worry about data quality?” The answer is, if all you are doing is following a CAD paper-based process, requiring no downstream uses, then the quality of the data is not critical.

However, if you are following a BIM process where data is being generated to support downstream uses and the creation of a ‘digital twin’, it is a very different story. The following is a link to a recent blog written by Dylan Underhill, Digital Twins – A Beginner’s Guide, setting out the process and clarifying what a digital twin is.

Defining Required Data Points

Assuming that use of the underlying data is a goal, then ensuring that the data is fit for purpose is extremely important. What steps need to be taken?

First, what are the required data points? Have they been defined? In an earlier blog, LOD – Level of Data, Level of Development, Level of Detail or just Level of Disorder?  I discuss the need to understand LOD and its limitations and the requirement for provision of a clear ‘Level of Information Need’ as defined by ISO 19650. Without this, there is no agreed definition of what is required, why it is required, who is responsible, when it will be available and, therefore, what is to be checked.

Once the data points are established, then to reduce risk and enable teams to use the data with confidence, the data quality needs to be checked, in the same way that drawing output is checked for compliance. The strategy for auditing the data quality will depend upon what it is that the data is required to do.

Good Quality Data

There are two fundamental ways in which the data will be used: data conversion (changing computer data from one format to another so another program can use it) or data migration (moving the data from one data base to another). For either data conversion or migration to occur, you need good quality data, meaning it must be:

• Accurate – correct, consistent, and precise.
• Consistent – same format and units, e.g., the date format is always the same.
• Complete – all required data is present by the agreed submittal.
• Valid – fields are entered correctly and in the correct format.
• Timely – current, up to date and available for immediate use.
• Accessible – understandable and easy to access.

Benefits of Quality Data

The aim of spending time and, therefore, money auditing the underlying data in a BIM process is to provide the following benefits for those relying on those downstream uses:

• Confidence in the information generated.
• Ability to use the data to improve downstream processes, generate reports, undertake variance checks, and have access to accurate information to support decisions.
• Spend less time, fixing, organizing, and aligning data fields.
• Users of the data are happy with the outcomes.
• Reduction in cost through efficient use and reuse of the data.

Risks of Poor Data

On the other hand, missing this quality step results in a risk of low quality or ‘bad’ data which can lead to the following challenges:

• Error through the use of incorrect data, incorrect conclusions are reached or discissions made.
• Duplication of entries through multiple points of entry, leading to a lack of trust and errors.
• Increased frustration leading to reluctance to use and rely on the information.
• Inability to upgrade the data to work with new technology, in is not ‘future proofed’.

Having established the need to audit the underlying model data, who does it and how is it done? As with any process, each consultant is responsible for the quality of their own data, not just the drawing output, and should have processes and protocols in place to check their own model data set. The design and construction teams should have addressed, within their BIM Execution Plan, how they are assigning responsibilities and managing duplication. There are several ways to do this and that is the subject of another blog.

In addition, as with any QA/QC process, an external, impartial audit is always a good idea. At Summit BIM we have developed robust protocols to review and check model data against different types of issues, dependent upon the defined downstream uses. We use our own software, BIMFMi^©, to quickly and easily isolate, sort, filter and organize the model data sets to ensure that what is required has been provided.
If you would like to learn more, please reach out to us.

Contact us to get started