Data science made a noticeable impact on the public sector in 2017. And it’s no longer just a small group of innovators who are putting data to work in their organisations. Projects such as the GLA’s datastore are helping the sector build a data infrastructure that will benefit us all.

While progress has been made, the pressure to deliver more services whilst cutting costs remains. To help meet that challenge, here are some of the new things we’re busy building for Witan, the hub for data-driven government.

Simplifying data governance so that GDPR is an opportunity, not a threat

The GDPR is turning the world of data on its head by taking control of personal data away from data “holders” and putting it back into the hands of citizens. Yes, this means that cities, government departments and local authorities need to tighten up their systems and comply with new regulations. It’s also an opportunity for organisations to get a better handle on data by managing it properly and putting it to use to get new insights, build new services and save money. But that takes effort. In the words of the National Infrastructure Commission’s recent Data for the Public Good report, the key is to collect high quality data and use it effectively. In late 2017 the first version of our Datastore product was launched to do just that (and is already supporting the GLA).

Here are some of the features we’ll be making available for Witan Datastore customers in 2018:

• Audit what happens to your data with access and event recording. Showing you’ve got an adequate system to manage and control private data is essential for GDPR compliance. Datastore will allow you to understand, track and report on everyone who has access to your data and where it has been used, with in-built analytics on usage and downloads.

• Share data, charts & scripts using flexible access controls. Collaboration is often the key to getting insights from data. Datastore will allow you to share data, reports, charts, scripts and files with partners, as well as internally. When it comes to private data, this is essential for the GDPR, as private data can only be used for the express purpose an individual has given you consent for.

• Explain what’s going on with data through metadata sharing. Open up more opportunities to drive value from your data by showing what data exists, and the latest progress, while ensuring that the data itself is not shared unless with your explicit agreement.

• Make data easier to find and use with editable metadata tools. This feature allows you to provide clear descriptions of data including structure, size, license, creator and location. You will also be able to communicate the methodological, ethical and legal uses for data.

• Share ideas, analysis and insights using “datapacks”. This is our way of making it easy to bundle up and share what you need when you want to present your analysis and modelling outputs. Scripts, code, charts, input, output and reports can all be included in a datapack to present analysis and modelling. This curation allows a team to share the right versions of data and scripts, along with the conclusions they reach with that data, making their work more transparent and making it easier to reconstruct or check later if needed.

Using models to generate new insights into critical public sector services.

Understanding demand for public services and making decisions about provision is a fiddly process. Getting it wrong can mean budgets get blown and people don’t get the support they urgently need.

Our first demographic model was built in partnership with London boroughs in 2016. In 2017 we launched the first iteration of our SEND model - we focused on building a tool that would simplify the complex process of service and budget planning for children with special educational needs.

In 2018 we’re adding further models to our decision-maker toolkit and further developing our existing models. Here’s an overview of what we’re building.

• A more comprehensive SEND demand and policy model. We’re currently enhancing the model to deal with the more complex scenarios involved in SEND planning, including out of borough placements.

• New models for Children’s Social Care and Adult Social Care. We’re focusing on social care for three reasons. The first is the essential role social care plays in our communities. In 2016-2017, in England, our local authorities received 1.8m requests for adult care support. That’s 5000 people a day wanting help. This brings us to the second motivation for looking at the area, the cost. In 2016-17 local authorities spent £17.5 billion on Adult Social care and around £11 billion on Children’s services. Together this expenditure accounts for a significant proportion of local authority budgets, which is what makes predicting demand, and planning the most effective ways to meet that demand, so important. Our new models will enable decision makers to rapidly test out different demand scenarios, then model the impact of different service provision strategies and the impact of new policies.

It’s an area where small process and timing adjustments make huge differences to budgets. Which is why being able to compare different options is such an appealing prospect for the people in charge.

“data-informed decisions are decisions that can be trusted, decisions that will stand-up to scrutiny.”

The final reason that data science can help in this area is its complexity. Accurately modelling the impact of subtle shifts in demand and changes to service delivery means gathering large datasets together, performing complex analysis and sharing results. By hosting models, data and outputs on our Witan platform we make the whole process more efficient and more secure, giving decision makers the opportunity to test out new ideas. Perhaps more importantly data-informed decisions are decisions that can be trusted and will stand up to scrutiny, allowing leaders to be more confident in their actions.

Making data collection and sharing fast, efficient and secure

In the public sector, gathering data together, whether it’s for analysis or for statutory reasons, is a painful process. Have you been in the situation where for substantial parts of your year individual team members spend most of their time chasing and checking data from internal colleagues and external partners? You’ll know it ‘s massively inefficient and an endless, if hard to avoid, waste of time. And that’s before you start checking that the data is structured in the right way or start cleaning it ready for use.

In the next few months we’ll be launching Collect+Share, a tool that will make gathering data and safely sharing it with other people fast and secure. Here’s what you’ll be getting:

• Peace of mind by controlling what’s private and what’s shared through secure data access. Collect+Share will offer data controllers mastery over who has access to data to make it easier to conform to data legislation, organisational policies, ethical standards and user expectations.

• Get the data you need, fast, in the format you need with smooth data acquisition. You’ll be able to easily request data from multiple colleagues and external partners. Then track who has and hasn’t complied. People can upload data in familiar formats like Excel, saving them time and effort.

• Be sure that the data provided to you is ready to use with automatic schema and format checking. This is an important feature, as you’ll get automatic checking of the format of submitted files, followed by reports of errors to both the uploader and requester. This will speed up the delivery of correct data to where it’s needed and reduce the communications overhead of working to fix files or acquire accurate data.

What are we missing?

We’ve prioritised the development of Datastore, our social care Models and Collect+Share as we feel they address some of the most pressing problem in the sector. If you feel there’s something else the team should be considering, or a practical application of data science that could make a difference to your community we’d love to hear from you. Email theteam@mastodonc.com