Despite the importance of quality assurance in ensuring data projects are accurate from conception to deployment, this is a process that many tech companies struggle to perfect.
The 2024 State of Testing Report revealed test cases are not well-written and maintained for 60% of organisations, highlighting the challenge tech leaders face to deliver a seamless testing phase.
According to Maksymilian Jaworski, Data Engineer at IT consulting company, STX Next, detecting and addressing inaccuracies as soon as they arise minimises the risk of propagating errors, simultaneously reducing the cost and effort required to fix them.
Maksymilian takes up the story: “In data engineering, the principle of ‘validate early, validate often’ emphasises the importance of integrating validation checks throughout the entire data pipeline, as opposed to deferring them to the last possible moment.
“Handling data quality issues at source is by far the most cost-effective method of operating. Dealing with unforeseen roadblocks during the remediation phase is significantly more expensive, while problems at the deployment stage can cripple a data engineering project. This underscores the value of implementing a rigorous quality assurance regime, that spots and eradicates any outliers early in the project cycle.
“Programming data transformations is a minefield of avoidable errors. Common mistakes include forgetting to add a required argument to a function, trying to access a column missing from a table produced upstream, or attempting to select from a table that doesn’t exist.
“Typically, a trivial solution is required to fix these issues – what’s crucial is the stage at which problems are discovered. Manually testing code can uncover inaccuracies at an early point in the data engineering process. Mistakes will also show up when code is deployed to production, but this is far more costly to fix. Although unit testing is recommended by many data engineering experts, this is often a laborious and unnecessary process that hampers further development.
“External testing of the application is another effective method of quality assurance. This is where the application is run in a simulated environment, with engineers checking that the results match the expectations of the given test case.
“Finally, tests should be put in place to ensure that the data supports business operations and decision-making. Organisations must guarantee the consistency, completeness, timeliness, accuracy and referential integrity of outputs, all while making certain the data adheres to specific business rules.”
“Data engineers must take a long-term view when it comes to quality assurance. Investing time and resources into running tests at the nascent stage of development can prevent costly errors further down the line, potentially preventing a project from being delayed or even scrapped.”
For more from STX Next, click here.
The post Tackling bad data at source is key, STX Next claims appeared first on Data Centre & Network News.
Despite the importance of quality assurance in ensuring data projects are accurate from conception to deployment, this is a process that many tech companies struggle to perfect.
The 2024 State of Testing Report revealed test cases are not well-written and maintained for 60% of organisations, highlighting the challenge tech leaders face to deliver a seamless testing phase.
According to Maksymilian Jaworski, Data Engineer at IT consulting company, STX Next, detecting and addressing inaccuracies as soon as they arise minimises the risk of propagating errors, simultaneously reducing the cost and effort required to fix them.
Maksymilian takes up the story: “In data engineering, the principle of ‘validate early, validate often’ emphasises the importance of integrating validation checks throughout the entire data pipeline, as opposed to deferring them to the last possible moment.
“Handling data quality issues at source is by far the most cost-effective method of operating. Dealing with unforeseen roadblocks during the remediation phase is significantly more expensive, while problems at the deployment stage can cripple a data engineering project. This underscores the value of implementing a rigorous quality assurance regime, that spots and eradicates any outliers early in the project cycle.
“Programming data transformations is a minefield of avoidable errors. Common mistakes include forgetting to add a required argument to a function, trying to access a column missing from a table produced upstream, or attempting to select from a table that doesn’t exist.
“Typically, a trivial solution is required to fix these issues – what’s crucial is the stage at which problems are discovered. Manually testing code can uncover inaccuracies at an early point in the data engineering process. Mistakes will also show up when code is deployed to production, but this is far more costly to fix. Although unit testing is recommended by many data engineering experts, this is often a laborious and unnecessary process that hampers further development.
“External testing of the application is another effective method of quality assurance. This is where the application is run in a simulated environment, with engineers checking that the results match the expectations of the given test case.
“Finally, tests should be put in place to ensure that the data supports business operations and decision-making. Organisations must guarantee the consistency, completeness, timeliness, accuracy and referential integrity of outputs, all while making certain the data adheres to specific business rules.”
“Data engineers must take a long-term view when it comes to quality assurance. Investing time and resources into running tests at the nascent stage of development can prevent costly errors further down the line, potentially preventing a project from being delayed or even scrapped.”
For more from STX Next, click here.
The post Tackling bad data at source is key, STX Next claims appeared first on Data Centre & Network News.