Data quality: measuring and improving

Data is the new gold, or so we've been hearing regularly for several years now. With gold, however, we know that its value varies with its quality. 22-carat gold is worth roughly three times as much as 8-carat gold. But how many “carats” is your dataset? How do you measure data quality? And what can you do to improve the quality of your data? We explain it all in the blog below!

Why is data quality important?

Before we explain how to measure data quality, we first need to consider why data quality matters in the first place. High-quality data is more valuable because it describes reality more accurately and is easier to use for different purposes. A common saying is “garbage in, garbage out”: if you use poor-quality data, you'll end up with poor-quality dashboards and insights. It's no coincidence that data quality is one of the top trends in the BARC Data, BI & Analytics survey of 2022.

How do you measure data quality?

But what exactly is data quality? Or rather, how do we measure it? A common mistake is to equate data quality with data accuracy. For example, we know that the first name of the Dutch king is “Willem Alexander”, so if we see “Willem Alexia” or “Wellim Elaxendar” in a database, we know that value is incorrect. Accuracy, however, is just one of the dimensions of data quality. Different organisations and data quality frameworks each have their own set of dimensions, but in general data quality is defined as the degree to which the dimensions of data meet their requirements. The tricky part of this definition is that those requirements can differ, and not all dimensions are equally relevant to every business or process. The international body DAMA (Data Management Association) defines no fewer than 60 dimensions of data quality.

Measuring data quality therefore starts with selecting the data quality dimensions that are relevant to you. DAMA has a 4-step plan for this:

1. Determine which of the 60 dimensions apply to the data whose quality you want to measure. If you're dealing with master data (often dimensional data such as locations, products or people), for example, different dimensions apply than when you're working with transactional data.
2. Next, determine the added value of each dimension for your business or process. Measuring, monitoring and improving data quality costs time and money, so the added value has to justify it. For one organisation, the degree to which a record complies with an industry standard may be very important, while for another there may not even be an industry standard. In that latter case, the dimension adds no value.
3. Prioritise the chosen dimensions based on costs and benefits. That way, you can tackle the data quality issues with the greatest impact first and keep your budget under control.
4. Define the KPI and measurement method for the dimensions you want to monitor and improve. At the same time, make sure you jointly agree on the target value for each KPI.

Primary dimensions

Of course, it's a good idea to apply the above to your own situation and choose the right dimensions. But to give you a better feel for what the dimensions actually mean, below we look at the 6 so-called “primary dimensions”. These dimensions apply to virtually every dataset and are a good place to start.

Accuracy

We already mentioned accuracy above: it's the degree to which data matches the real-world value. Examples of errors that can occur here are:

• A person's name is misspelled, for example Janssen instead of Jansen
• The unit of the input is incorrect, for example the value 200 for the weight of a bag of sweets in a column where the value should be entered in kilograms (instead of grams)

Completeness

Completeness is the degree to which all values and records are actually present in a dataset. This dimension can go wrong in two directions. Values can be missing, but there may also be values present that don't belong in the dataset. Examples of errors include:

• Not all products that are sold appear in the products table
• The customers table contains people who aren't customers
• For some products, the colour column hasn't been filled in

Consistency

Consistency is the degree to which a value in two datasets, or within one dataset at different points in time, complies with specific rules. Examples of (in)consistency are:

• An address is registered in Amsterdam, but with Germany as the country
• A customer has one address from 1 January 2019 to 31 March 2022 and a second address from 1 January 2022

Timeliness

A dataset is timely when it is available within the agreed timeframe. For example:

• The previous day's orders must be available by 07:00 at the latest, but they could only be viewed at 07:30

Uniqueness

In principle, each value in a dataset should occur only once. When a value occurs more than once, it's a duplicate. This can happen in different ways:

• The entire record occurs twice in full:

• The same value occurs with 2 different keys:

• The same key occurs twice with different values:

Validity

A value is valid when it complies with the rules that apply to that type of value. Note that a value can be valid without being accurate or complete. There are several ways in which a value can be invalid:

• An email address without an @ in it
• A Dutch 06 mobile number that doesn't consist of 2+8 digits
• A date of birth in the year 2202

Improving data quality

Once you've chosen a set of dimensions and defined the metrics (KPIs) for them, including the associated target values, you can start systematically monitoring and improving data quality. When you begin this process, you may quickly discover that your current datasets contain all kinds of problems that cause a low score on your data quality KPIs. The key is then to solve these problems systematically. The best approach is the “leaking tap” method. Here's how it works:

Imagine you walk into the bathroom at home and suddenly see a big puddle of water. You have a leak! The first thing you do is track down the cause of the leak. You soon see that someone didn't turn off the tap, causing the sink to overflow. You could start mopping, but that's pointless, because water keeps flowing from the tap. So first you have to turn off the tap. After turning off the tap and mopping up the water from the floor, you could walk away, but then there's a risk that someone else will forget to turn off the tap again, so you decide to think about preventive measures. By fitting an overflow in the sink, you prevent the leak from happening again.

We can apply the same approach to a data quality leak. Suppose you come across all kinds of variants of a bottle of “water 500ml” in a database: “water”, “water 50cl”, “water spa blauw”, “waater”, etc. This is an example of an accuracy error we want to fix. Just like with the leaking tap, we don't start by writing queries to change all those different values to the desired “water 500ml”, because as long as water is being sold, the incorrect values will keep coming. Instead, we walk over to the cashier and ask her to always enter the correct value from now on. Then we can run our query to correct the incorrect values. Finally, we go to our IT department and ask them to change the free-text input field into a drop-down menu from which the cashier can only select the prescribed values. That prevents her from still making a spelling mistake, or a new cashier from making the same error, and the improved data quality is guaranteed.

E-mergo can help

In this blog, we've explained the basics of data quality. It's important to realise that good data quality isn't a one-off project, but a continuous process. Safeguarding good quality requires an interplay of people, processes and technologies. E-mergo can help you with this process in several ways:

• With low-code applications built in Mendix or Power Apps, we create apps that guarantee high-quality input.
• In TimeXtender, you can define data quality checks that run automatically whenever new data enters the system.
• With Qlik Data Integration, we can process data in real time using Change Data Capture, so you always have timely data at your disposal.
• Dashboards in Power BI or Qlik can give you insight into your data quality KPIs at a glance, so you can tackle new data issues straight away.
• With data strategy workshops, we can help define the right processes to continuously monitor and improve your data quality, and together we determine the next steps for your organisation.