Data-led decision making is the holy grail of government delivery; finding it hinges on the ability to provide administrative data and statistics faster, more frequently and more effectively. This will require moving away from old ways of working, says the Scottish Government’s Chief Statistician, Alastair (Ally) McAlpine.
McAlpine was appointed Interim Chief Statistician for the Scottish
“What I would like to see happen across the public sector in Scotland is that we have got data-driven policymaking and that the services we're providing are based on data evidence,” he tells Government Transformation Magazine. This will involve working with about 30 different organisations that produce statistics for official use and aligning them on a clear vision.
McAlpine has first-hand experience of harnessing the power of data for public good having led the Scottish Index of Multiple Deprivation in a previous role, which is government’s official tool for identifying deprived areas across Scotland. He’s also co-director for Administrative Data Research Scotland, where he is helping to make data for public good more accessible to researchers.
There are familiar challenges when it comes to data-led decision-making; from finding, navigating and interpreting data, to analysing and distilling it into compelling insights. On top of that is the pressure to do it faster and more frequently while retaining public trust and confidence, McAlpine explains. “There's a lot of good places where this is happening, but we need to speed up the time it takes to get good quality data and improve the value and relevance of our statistics and analysis."
With the numerous different ways in which administrative data is being collected across the public sector, McAlpine says he wants to streamline how data is going into government while reducing the variability and the quality of the data that's coming in as well. This means ensuring there is a methodology around what information is being collected, in what format and why.
“We need to move much faster on Reproducible Analytical Pipelines, so being able to think about the analysis we're producing. This is key to building public trust in our data analysis and transparency around how we derive statistics.”
There’s also a need to improve the consistency in the approach to data across government, McAlpine explains. “Too often the benefits of using data for tangential work in government and across the public sector are not front of mind for the person who's responsible for managing the risks.”
Thinking back to just over two years ago, during Covid-19, everybody knew the benefits of what data could do. Now that we've moved out of that emergency environment where we're constantly responding to a crisis, McAlpine argues that those benefits and risks are stacked separately again. “We need to start thinking about how we can bring them back together. It’s not impossible to overcome but it has created a drag with regards to using data quicker. It’s something I want to be tackling as Chief Statistician.”
There are over 300 statisticians in the Scottish Government, but as McAlpine notes, there’s been a lack of training aimed at developing leadership skills for this group of professionals – something he is determined to change. Over the last year, work has begun on developing a training course called ‘Fit For The Future’ which focuses on equipping statisticians and analysts with the project management tools and knowledge of change theory to take data leadership forward.
“What I'm trying to do is improve the scalability of change,” McAlpine explains. “As we transition into a modern world where data is flowing, we need to be able to adapt – that takes strong leadership. We need more pathways to meet those data ambitions and we need data analysts to have the confidence to lead on some of those changes. "
McAlpine admits that this is the project he is “most excited” about. “Often the onus is on me as leader to talk about the change I want to take forward, but the most important factor in all this getting other statisticians to be thinking in that space as well.”
In preparing for a government in which data will play an increasingly crucial role, McAlpine acknowledges the need to embrace new ways of working. “Within statistics, there's a lot of traditional ways of working but I think we need to embrace different ways of doing analysis. We also need to be pragmatic about this, accepting that some things we might not always get right. But we need to keep pushing the boundaries and see what else we can do.”
This will require moving away from top down management styles of leadership which is traditionally seen across the Civil Service. Giving people the freedom to explore, McAlpine argues, is the key to an agile and responsive organisation. “There's a generation coming through that wants to change the world and want to make it better. As a leader in the Scottish Government I want to make sure that we can harness that energy going forward.”
With the huge boom of generative AI throughout last year, how the information the public sector collects is being used has been flipped on its head. “You can imagine the world where future policymakers might be speaking to an AI model instead of an analyst so we need to make sure we've got data in the right place,” McAlpine says.
At the moment he’s focused on taking advantage of other AI and Machine Learning tools. “I think neural networks are still really important and I think random forest classification methods will continue to be incredibly useful.”