ODI-Cardiff is working to strengthen the data ecosystem across Wales. One thing we’ve heard from loads of people is that they’d like to meet other data folk in real life. But we’re all busy people who have to earn money and there’s a limit to how many meetups you can go to in any given week.
So here’s the plan…
We’ll all get together once a month. And do real work.
Thanks to our lovely friends at IndyCube we have access to co-working space in Cardiff (right by the station).
So on 26 September you can just turn up with a laptop (or some paper and pens, we won’t judge) and do some work. And lots of other people will do the same thing. And we will all be people who work with and are interested in data.
This isn’t networking, this is casual co-working. No-one will hand round business cards or give us their 60 seconds pitch. You don’t need to take time out of your work. You simply come and work with us for a day.
The wifi is rock-solid, the chairs and desks are comfortable, the coffee is surprisingly good. You might even strike up a conversation over coffee. Maybe this will be the start of a collaboration that will change the world. Maybe someone will show you how to make Tableau that thing you did once by accident but never managed to repeat. Maybe you could just benefit from some different scenery than your normal office for a day.
You don’t have to work with (or even care about) open data. Any sort of data is fine. You could be a developer, a designer, a data-scientist or an enthusiastic policy wonk. You might work for yourself or the government, or the man, it matters not to us. As long as you love data.
A small social enterprise called Democracy Club develops tools to make it easier for UK citizens to exercise their vote effectively. One of these tools is “Where do I vote?” Give it your postcode and it tells you which polling station you should go to. Where do I vote? is apparently one of the most Googled terms on UK election days. They have an elegant tool but they need the information on which addresses vote at which polling station. In Wales this information is held by local councils. Democracy Club asked all Welsh local councils to provide this data in time for the General Election in 2017.
How did Welsh local authorities do?
Fourteen councils provided data (green in the map), eight councils did not (red in the map). The same data is in a table below. There were some problems with the Cardiff City Council data and a few specific areas of other councils.
Overall the 92,000 searches of WhereDoIVote? made for Welsh postcodes, 51,000 resulted in people being told their polling station. Which is nice for 51,000 voters but disappointing for the 41,000 people who didn’t get a result.
This is a good litmus test for how local authorities approach open data because:
Democracy Club is after a specific dataset, they understand the data and can help with any questions council staff might have (they make it easy for the councils)
if the council provides the data it will directly benefit their citizens
the council definitely holds the data (it needs it to send out the polling cards that so many people apparently mislay by polling day)
So if councils can’t give Democracy Club this data, it is a signal that they have quite a long way to go in understanding the power of opening datasets. We’re interested in developing a scorecard for Welsh local authorities in terms of open data. Should we include providing this data to Democracy Club in that scorecard do you think?
“to enable the Welsh public sector to collaborate more closely in procuring goods and services”.
The NPS is in the midst of supplier engagement on a new framework to cover digital services. Given the existing UK frameworks for digital services this may or may not be a good idea. NPS feels the UK frameworks don’t address language issues and other requirements particular to Wales
The Welsh Government is “Committed to Open Data” according to its Open Data Plan. As far as we can see the NPS has no open data plan and the work on the digital framework is not considering the government’s open data plan at all.
This is a pity because open data and procurement are perfect partners. Open data can drive efficiency and transparency in procurement processes as well as having wider benefits.
A fit for purpose procurement process should address open data at least in the following areas:
performance against contract targets
Suppliers should be required to publish their performance as open data. And the performance of previous suppliers should be published as open data. This will help customers and suppliers understand which are the challenging parts of the contracts and should lead to improved performance over time.
administrative (exhaust) data
In order to deliver a service, lots of data has to be collected and used. This data is typically locked in the systems of suppliers but if it is published as open data has potentially high economic value.
In order to deliver a service, a supplier may need to collect data for reference (to provide IT support a supplier will need to know where local authority buildings are). Ideally suppliers should use existing open sources of this data and they should certainly publish reference data that they create in the contract.
Many contracts let by public bodies in Wales have as their primary purpose the creation of data or information. These datasets should be open by default and published to a high standard.
We really shouldn’t be having to have this conversation in Wales. We should be talking about how quickly we can move the open data to five star status. We should be talking about what the open data tells us about public services in Wales. We should be using the open data to help us understand the impact on the wellbeing of future generations.
Let’s hope we can start having those conversations soon.
Transport for Wales consulted people on the design of rail services for Wales and the borders. ODI-Cardiff responded to that consultation. This is a copy of our consultation response. We work with and support the network of people working with open data across Wales. This consultation response was put together with the assistance of many people from that network.
The new rail franchise presents significant opportunities to benefit the communities of Wales and the borders. The consultation on the design of the rail franchise leaves out a vital component. If TfW addresses this, it will significantly enhance the effectiveness of rail and other transport service design.That component is data.To operate effectively, the franchise holder, TfW and other stakeholders will need to share a wide range of data across the life of the franchise. The franchise will also generate data through its operation and will consume data from third party sources.The decisions taken by TfW at the start of the franchise about the publishing, sharing and ownership of data will have profound consequences on design possibilities along the franchise life.As a hypothetical example:In the future, metro rail services may be tightly integrated with bus services. As a light rail train is ready to depart a station it would be helpful to understand the location of a bus that is delayed and the number of passengers the bus is carrying. If the bus is nearby and is full the rail service might wait for the bus. This prevents overcrowding on the platform and significant ongoing delays to following services.In order to achieve the scenario described above would require good data infrastructure, combined with dynamic timetabling and effective automated decision making. Those things do not need to be designed and built now. They may never need to be designed and built by TfW or the franchise holder at any point. As long as the data infrastructure is in place, then the technology that uses that data can be supplied when it is appropriate and, potentially, by other investors in the technology market.If TfW specifies a largely closed data infrastructure (or worse still, does not specify data infrastructure at all), then little data will be available to third parties. It will be hard (and potentially expensive) to gain access to data for those third parties and innovation is likely to be significantly hampered.If TfW specifies a largely open data infrastructure, third parties would be able to access relevant anonymised data dynamically via APIs and to use and reuse the data in their own services. This will support innovation in services and access to services across Wales and the borders. This innovation could come at no cost to TfW or the franchise holder and without requiring those bodies to become involved in the development of services.Network Rail is already providing such open data to the community on a national basis leading to websites such as Real Time Trains. The decision by Transport for London to open up their data led to over 8,200 developers signing up to the program and more than 500 apps being created in the first two years that it has been online.Open Data and associated “Hackathons” have provided innovative solutions. For example Transport for London has been able to provide better quality service to their customers whilst reducing the normal costs associated with developing such software.For these reasons we strongly encourage TfW to urgently develop a data policy and infrastructure plan. We propose a set of principles which, we believe, should underpin such a policy and plan if they are to be effective for Wales.
Our proposed data principles for the rail franchise
TfW should consider at design stage the rights in the data that is used and created across the franchise. Some of these rights will lie with the people using the service. These rights are expanded in the GPDR.TfW has an opportunity to use data infrastructure to give people more control over the data about them that will be gathered within the franchise. TfW should use open APIs to create opportunities for innovation by the people and companies of Wales and the Borders.For the avoidance of doubt TfW should be clear about who has rights over what data across the franchise.Where possible data created as a result of the operation of the franchise should be vested with TfW. TfW should also have rights to all data inputs into the franchise unless it is not legal or feasible to do so. Where TfW cannot have rights to the data it should by contract ensure that data is published and shared in a timely fashion in line with Principle 1.
Principle 3: data infrastructure plan
TfW should develop, publish and maintain a data infrastructure plan.This should identify the datasets that are required and will be generated within the franchise. The plan should also show how that data will be governed in a way that benefits the communities of Wales and the borders. There is more information how governance relates to data infrastructure for transport in this report for the Transport Systems Catapult.It is important that this plan is owned and maintained by TfW itself and not by the franchise holder to avoid conflict of interests over the collection and release of data.The data infrastructure plan should make it clear how datasets will be maintained and accessed by stakeholders. We suggest that TfW should minimise the requirement to hold data itself. Instead it should require data to be made available to itself and to other stakeholders via suitable APIs. Given the length of the franchise and the speed at which digital technology advances the data infrastructure plan is likely to require significant redrawing at regular intervals through the life of the franchise.
Principle 5: develop policy about withholding data for security or crime prevention reasons
TfW should develop, in partnership with relevant experts, a policy toward the suppression or masking of data for security reasons or the prevention of crime. This investment in policy would be to avoid unnecessary or unhelpfully risk averse decisions being made during the life of the franchise.Any decisions not to publish, share or open certain datapoints or datasets for the protection of the public should be made against a clear policy drawn up with a clear understanding of the risks and benefits of publishing and suppressing certain types of data.Where possible, and in the interests of transparency, the reasoning behind each decision should be placed into the public domain to ensure that all and any users of the system understand the restrictions on the datapoints/sets. The police already publish crime maps and other forms of open data and may have advice to contribute.
Principle 6: use of third party data
TfW should lay out its policy towards the use of third party datasets. In particular throughout the franchise there should be a presumption not to duplicate data held by other organisations but instead to access the data when required via APIs (for example accessing crime data through data.police.uk). TfW should also encourage third party holders of relevant data to publish their data under open licences via APIs and adhering to appropriate open standards.
Principle 7: working with open data communities
TfW should encourage the use of open data sources wherever possible and actively contribute back to these datasets.
We hope all political parties in Wales will follow her two simple suggestions in the run up to the election and so we have republished her letter here and sent a copy the main parties in Wales.
The UK will have a general election on 8 June 2017. You and other political parties are putting forward candidates, creating a manifesto that says what you will do if you form part of the government after the election and campaigning for the votes of the electorate. If you form part of the next government, you will want to deliver your manifesto and policy ideas.
Like every other organisation you will find that in the 21st century data and technology are essential tools to help with all of these things.
Voters need to know what your policies are, who your candidates are and what those candidates think. To get this information to people you will need to collect, use and publish data online, and you will want people to trust that you are doing this in a way which is ethical. Meanwhile, if you get into power you will find that data and technology create new ways to deliver your policies in ways that meet the changed expectations of citizens.
Unfortunately, recent political campaigns have had common themes:
concerns over the use of personal data and targeted advertising
electoral processes that have not been adapted to the 21st century
a lack of debate over the changes to our society that are being brought about by the internet and world wide web
We are asking you today to make two simple enhancements to your practices in the run-up to the general election:
Publish information about candidates as open data
Openly state how you use and reuse personal data in the cause of your campaign
Our goal with this request, as with other things that we do, is to encourage the best use of data, to build people’s trust in how organisations use data and to help people make informed decisions. In this case, the decision people are making is who should represent them in our democratic system.
We are here to help and to work with you. Please email firstname.lastname@example.org if you need support or would like to discuss this further.
Open Data Camp was held in Cardiff a couple of weeks ago. It is a free gathering of people from across the UK and further afield who are interested in publishing or making use of data. Especially open data.
Ben from the ODI-Cardiff core team pitched a session on the target the Welsh government has set to increase the number of Welsh speakers to 1M by 2050 (roughly a doubling compared to today). The discussion focused on what forms of data and what models are available or could be developed that would help inform and guide policy at a national or local level to meet this target.
A live blog of the session and collaborative notes are available (in English only) so we won’t repeat what was covered here.
We have a project
We will be doing some work on this within the ODI-Cardiff community over the next few weeks.
What we definitely plan to work on includes:
Taking the population growth modelling for Wales as a whole and for local authorities (from StatsWales) to look at what seems likely to happen to the language with no further interventions
Looking at the population growth models in more detail to see what sort of interventions in schools would be necessary to get close to meeting the targets
Looking at the local authority data to make estimates of what each local authority might reasonably need to achieve in terms of Welsh speakers to contribute to the 1M target over all
We would like to work on things like:
More sophisticated modelling based on an understanding of how levels of speaking other minority languages have changed in other contexts
Building a model of Wales and language use so that the likely impact of policy interventions can be predicted and policy can be improved up front
Staff from the National Assembly took part in the session and are keen to see what the community can do with the data that is currently out there.
We’d really like your help.
We especially need people with modelling skills (in any field not necessarily Welsh language models) and people who know about language use and change.
That said, enthusiasm is the key skill. If you’d like to get involved, please get in touch now.
(We don’t have any funding for this, this is strictly a volunteer and community project).
The latest Digital Tuesday event focused on The Open Data Challenges for Wales.
ODI-Cardiff was ably represented by Esko Reinikainen.
During the talk he covered the mission of both the ODI and ODI Cardiff, data as infrastructure, the data spectrum, recent projects from Wales, and invited everyone who is interested to join our online community. There was also a call to help us map the open data ecosystem in Wales using a one minute survey. If you are a data producer, publisher or consumer in Wales, we would like to hear from you.
We’d like to highlight a nice piece of data journalism from Ceredigion.
Cambrian News reporter Caleb Spencer wanted to make the information on councillors’ attendance at meetings more accessible to his readers. He created a visualisation in CartoDB which breaks the information down by ward.