The Department for Work and Pensions’ Sheffield Data Science Hub is working on a plan for a service harnessing data to monitor the performance of pilots of digital services.
Head of the hub Tetyana Mykhaylyk outlined the concept of the digital trialling framework at UKAuthority’s Data4Good conference last Friday.
She said it is hoping to begin using the concept within the department over the next few months, although the exact schedule depends on teams being ready to do so for their projects.
“Our idea has been to develop a robust mechanism of collecting data, transforming it into features and providing information on a trial’s performance,” Mykhaylyk said, adding that it would be in near real time. This would support an agile approach to running trials and making the relevant business decisions, extending to the ability to halt a trial if it was not running to plan.
She said the team has created a robust, semi-automated and repeatable data pipeline to produce datasets with features agreed with the teams using the service, and a trial monitoring tool to visualise measures of success.
This should help its users understand how their trials are progressing, largely by moving it to the second stage of the ‘wisdom pyramid’, in which data is transformed into information. (The upper stages are knowledge and wisdom.)
“Data is often unstructured and complex, but by going through this pipeline we can provide additional information, so we can structure the data and give our customers information” she said. “There is plenty of data, but the problem is what do we need from it?
“We need to spend a lot of time in understanding the data, especially if it is unstructured, but we think this can be the new agile way of monitoring trials in the future.”
Image from justgrimes, CC BY 2.0 through flickr