The British government recently announced an expansion of its Data Science Accelerator programme, aimed at training civil servants’ in data science skills.
The scheme is part of the government’s efforts to improve the way it uses the vast amounts of data it holds, a programme which also includes moves to make it easier for departments to share data through new rules set out in the Digital Economy Bill˚, currently making its way through Parliament.
All of this is welcome. Data is an increasingly vital tool for all large organisations, and the public sector is no different – if anything its public service role combined with access to rich personal data which we would rightly deny corporations, compels it to do more to benefit its citizens.
It is good to see that the government is recognising the potential of this data and looking to extract the value that lies within it. The public sector has data in its hands that, if used correctly, could improve operations and services exponentially, from managing transport to spotting tax fraud. Huge savings could be made to the crisis stricken NHS through better use of data.
The government has taken a very wise step in ensuring that staff have the capability to understand and utilise the data it holds. Training staff will have long-term benefits as it looks to get more value from these huge pools of information.
But training staff to be ‘data savvy’ is one thing. Training staff to be data scientists is quite another. The suggestion that a course in data science of one day a week for three months will create a workforce of experts is unlikely to be the case.
Getting everyone to record data in the right format, spot particular patterns in data, or to interpret visualisations is all very useful. But the danger is that it ends there and the real, transformational, national budget changing, life saving, projects don’t happen. The government ends up saying, ‘we’ve got some data scientists on it, they’ll find anything useful in the data, we sent them on a course’.
Is government underestimating the task?
Becoming a data scientist takes years of study. Becoming a good data scientist takes many more years of applying your hard-won mathematical skills to real world problems. But with this battle-hardened experience, great things can be done with the right data. And in the case of public sector data, we’re talking about the kind of things that can really change the way the world works.
I am not any way disparaging these courses or their usefulness. The work of the data scientists is far more valuable if everyone in the organisation gets what they are doing and is engaged with it. Nor am I saying that people working in government couldn’t become data scientists – I encourage them to do so, we need more. But it won’t happen in three months.
In the meantime, if the government is keen to really take advantage of all its valuable data, and get the ball rolling on analytics projects, it may need to take a different approach, at least in the short term, by bringing in established data scientists.
Even this is not quite enough. Data scientists on their own have a habit of just playing with the data. Government – like any organisation – must also have a clear strategy in place for extracting value from data, and agreement on what that the project should achieve. A department could have pools of data and access to the best data analysts in the world, but if these people don’t know what they are looking to achieve with this information, the project is likely to fail. As long as there is a clear understanding of what is to be achieved, this could be an effective technique to get fast results.
The amount of data that the government and public sector as a whole has access to is only going to increase and its value could be massive if exploited in the correct way. And there are certainly excellent projects already happening under the stewardship of excellent managers and data experts. These include HMRC’s Connect Programme using multiple data sources and analytics to track down fraudsters, through to making government data available for innovative web based platforms like NCUB’s Konfer service, which intelligently match businesses with challenges to UK academics who can offer solutions.
It is really positive to see that the government is engaging with the data revolution, and expanding the Data Science Accelerator programme is a great move. But the public sector has so many exciting things it could do with its data, I just hope it doesn’t end there with someone assuming that the data science box has been ticked. Having people with core data skills should produce incremental efficiency improvements, and is well worth doing. But big transformational projects need to be clearly defined from the top, well managed, and involve experienced data scientists. A data courses for non-specialist employees helps the success of such projects, but alone it does not deliver them.
Matt Jones is lead analytics strategist at Tessella