Southwark Council - Digital Futures Month Talk

Updated: Dec 13, 2018

“Find out about the ethical use of data to inform decision making and how it can solve real world problems for real people”


The brief:

  • Data and Analytics

  • Learn and be inspired by initiatives within the council and other organisations and see how to use amazing services driven by the smart use of data.

  • Smart Cities and Human Centred design.



Change mindset


In the next hour I hope that I can help to switch your Mindset and make you all want to do something different. With your help I want to take you all to an expert level in an hour.

Solving problems with and for the people in our borough needs us to change our mindset.

My mindset has changed since I have been able to meet lots more people and a wider range of people - having stepped out of local government for a while. I have done the same journey up and down the Piccadilly line for 15 years. This year I’ve got out and about - it’s been a year of saying ‘yes’. I want to bring some of these insights into this room.


Reflecting on Smart Cities - for me a city is not smart if it excludes people - makes the rich / poor divide worse. And it’s not smart or even civilised if we ignore people when they ask for help.


In fact - with data - perhaps we can even anticipate a situation and offer services before people need help? This is predictive analytics.


I am very much in favour of using anonymised data sets teamed with commercially available data and sensor data to find out, for example, which school playgrounds have dangerous air quality - for example one just near me in Bowes Park right on a bottleneck of the North Circular. On the other hand - is there a solution for this beyond relocating the school? Data can be like alchemy - uncovering invisible issues.


For me, to tackle the hardest and most expensive problems to society it’s not just about place but also people. Not just Geo but also Anthro (human). The only reason to capture people’s personal information is if we are going to do something useful with it. If we are going to take action and solve a problem.


It’s harder to work on people problems because we need to link up data sets to get insights at the same time as protecting people’s privacy.


In local government we want to help because we are compassionate human beings. And we need to do this because our jobs now involve being Public Sector Alchemists - we all need to come up with new ideas to save money and create better services.

Where do new ideas come from?


We needs lots of new ideas and that is where you come in. And we need to broaden our horizons.


This is what human centred design is about. The more ideas the merrier. And then use some tools even games - to prioritise which ones to action - democratically. Don’t throw away the rest in case the ones you have prioritised don’t work.


Where can these problem solving ideas come from? How else can we seed innovation or even productivity? We can start with some research data - which is exactly how to go about designing Smart Cities.


For me - the key thing that public servants need to do is to look much wider than ‘innovation teams’ to get ideas - before honing solutions down. We discount and disregard ideas much too soon because they don’t fit with our groupthink. Can we be more open minded?

The mindset we need is experimental, hypothetical “maybe…..”.


Front line staff that have been into people homes or talked to thousands of people on the street, will almost certainly have an in-depth understanding and some great ideas The service users themselves can bring their stories too.


Has there been too much focus on quantitative data - the bean counting and the performance stats - the SALT returns ‘Providing data on council activity’ rather than impact or outcome?


What we want to know is if people trusted us with their personal information and life stories and asked for help - did we do anything to help?


These insights - qualitative data - focussing on people’s experiences and stories - are equally objective. But they can be dismissed as anecdotal.


And right now - even with the data we already have - is there anything we could do to help - so people don’t have to tell their stories over and over? So that they don’t even have to ask for help but are automatically given the services that they are entitled to?


The answer to this is yes - and some of the front runners in Local Government data are showing the way. The good news is that here at Southwark you don’t need to reinvent the wheel. There is no shame in being a fast follower.


Practical examples


Here are some examples of this predictive approach and services already being delivered here in London and elsewhere. These are projects that I have worked on.


Camden


In Camden we pulled together 16 of the council’s biggest data sets and cross referenced them - both the people and the address data and from that the whole household. We did all the information governance to risk assess whether aggregating all this data was ethical. There is a real world risk of not sharing data - as well as the virtual risk of sharing it.


[SPREADSHEET TABLE]


This was in real time using artificial intelligence tools - a million records cross referenced overnight. This is about Artificial Intelligence as a productivity tool because people cannot do this in real time. In banking and retail this cross referencing of data sets is already well established - for anti fraud, money laundering and single views of customer for example.

It’s basically the same as breaking rocks - why would you do it with a mallet when you have a rock crusher / compactor?


Before that - for the troubled families programme in Camden it took a team of people 12 weeks to cross reference 7 spreadsheets. By then it was 12 weeks out of date.

This is about using the council’s own data - and getting value from it - in real time. With the residents index - a simple cross reference like an old-school rollerdex, Camden is able to:

  • See at a glance all the services involved with a child protection case, and see if any adults they don’t know about are living at the property. This was a key issue for the Baby P (Peter Connolly) case. Front line staff need up to dat intelligence to make the best decisions.

  • Save over £50,000 per sublet council property

  • For individual voter registration Camden proactively added young people to electoral register - and using local data matching - increase the % from around 40% matched to DWP records up to 80% using the residents index.

  • Proactively issue around 80% concessionary travel passes and blue badges rather than ask people to apply.

.

The London Office of Data Analytics


I was seconded from Camden to help Nesta - the innovation foundation - pull together data to tackle unlicensed homes of multiple occupation across London.


The idea was that - rather than random tip offs - we would use the Electoral register and Council Tax counts of residents at a property, combined with some open data property attributes on the numbers of stories etc. to produce prioritised lists for council inspectors to visit. We started with 12 but ended up with 7 boroughs that were able to contribute useable data in the timescales we asked for.


What was very disappointing was that a number of councils still do not cross reference their major systems to a standard address base (the technical term here is the Land and Property Gazeteer) so for this time limited project we could not link their household data with commercially available property attribute data.


This is basic housekeeping that should have been happening since the year 2000. Councils performance manage whether their major 15 systems are regularly updated (or better still - have an interface) to the LLPG.


And - more to the point - if you haven’t done the basics this means that when staff are out on site - if they have mobile devices - any new build addresses will not appear on their list. What happens next is that they make their best guess and type in the address - this is unlikely to be correct - and it will make lots of work down the line in data corrections and wasted visits to the wrong addresses.

So this was perhaps a failed experiment but - so long as you learn from it - it’s not a waste. The London Borough of Sutton have done some fantastic work on this since.


These were the findings:




[TECHNOLOGY, DATA, PROCESS, LEGAL DIAGRAM]


This is the data cycle to show how data can be used to speed up service redesign and have good quality data that people trust - to act on in emergencies:


[DATA AND SERVICE REDESIGN DIAGRAM]


I have also worked on the Worcester Office of Data Analytics which is a multi-agency project across the county - which I can answer question about if people are interested.


Finally - a sophisticated predictive analytics approach from the USA


Thorn is a data informed initiative to track down child victims of sexual exploitation and trafficking. After they have exited the abuse - some of the young people are surveyed and that information is used to design new interventions - truly agile -because the perpetrators change their methods all the time. On the basis of their experiences, the project works with commercial organisations (such as Craig’s list) to get data to help them find further victims. This is an excellent example of human centred design - a researched based approach using real life experiences to design interventions.


Brent Council - the forerunners in breaking down personal data silos since 2013 - are now doing something similar.


Here is ‘how to’ get started

[FOUNDATIONS FOR...SLIDE]


Bringing you up to an expert level

Councils have 800 lines of business and at least 400 database silos.


The beauty of being a fast follower council - as Southwark perhaps is - is that you can learn from others. Here is how most councils are doing data analytics - which in my view - is not sustainable and requires you to start from scratch for every use case.


Here is a question which needs data from more than one legacy system.


‘Is it the case that children who live further away from schools are more likely to be overweight?’


If we link the silos of data over and over again for every question (a ‘use case’ approach) then we are going to run out of time.


Alternatively here is a sustainable platform approach which is the route that Brent and Camden have gone down.


It requires an indexing level which enables silos to be linked and questions like the one above to be answered on the fly.


And if you need any more incentive -if you don’t take this approach and - under GDPR - someone ask for their personal data that you hold to be erased - you can’t comply - because you don’t know what information you have.


For those of you that are interested here are some further resources / developments

Designing for public services a practical guide

Information Sharing Gateway