Making the most of data: making data more open
A while back, I published an article on the use of open data sets, covering some of the things that were already being done and what may be possible in the near- to mid-term. After this was published, David Patterson of KnowNow Information (KnowNow) got in touch to ask to meet up to discuss what his company has been up to in this space.
I met up with David at IBM’s Hursley Laboratories in Hampshire. As a member of IBM’s Global Entrepreneur Programme, KnowNow has access to IBM’s skills and capabilities, helping to maximise KnowNow’s capabilities in how it deals with data.
With access to IBM’s Bluemix and Node-RED technologies, KnowNow can focus on its own deep domain expertise – managing datasets to ensure that its customers get the end results they need. Indeed, KnowNow has a viewpoint that is quite refreshing – it wants to promulgate what it is doing to as many people and companies as possible: it does not want to be proprietary. It wants to be able to monetise its domain expertise in how it understands, analyses and feeds information back as the customer needs it – not from dealing with data in any ‘hidden’ manner.
Part of this is undoubtedly based on KnowNow’s relationship with the Open Data Institute (ODI). The ODI is pushing for as many datasets to be made available via open APIs as possible. KnowNow’s approach has been to utilise as much ‘free’ data as possible – and therefore, it doesn’t believe that it can charge for the data itself – only for the results.
So, what has KnowNow been up to? The reason that David wanted to meet up was primarily to show me what KnowNow has been doing around flood monitoring and event prediction. With a system aimed at the Environment Agency and DEFRA in the UK, the idea was to be able to run predictive simulations of where resources would be required in the case of a flood event occurring. These resources could be anything from improved signage, presence of fire brigade, ambulance or army, based on available open data sets, included near real time data on rainfall, river levels and weather forecasts against geospatial information such as 2D and 3D map information. An example of an event in this case could be where a ford is likely to get to a level where a car could get washed away – the simple provision of a ‘no entry’ sign would prevent this. The use of the open data sets enables KnowNow to predict such events to a good degree of accuracy. It is easy to see how such a model can be used in areas such as brush or forest fires, drought and other weather events.
However, it is finding councils and central government hard to persuade to pay for the service – essentially, government is reactive, rather than proactive, so KnowNow may need to wait for the winter and for floods to happen before the government purse is opened. There has been more immediate interest from insurance companies – they can see value from this type of approach in setting premiums and in dealing with any fall out after an event.
KnowNow is also looking at other areas – and this is where its existing play in the open data market is key. The rise of the internet of things (IoT), could have a big impact on an areas such as healthcare. Take as an example a person who, for whatever reason, is still capable of living alone but requires a degree of oversight. At the moment, this will be carried out via timed visits from care workers – and it is apparent that the system is overstretched. Instead, create an intelligent environment around the person. Have they opened their medicine container when they should? Have they opened the front door at all in the past hours or days? Are they moving around? Has there been a sharp movement monitored as they move around, which could denote a fall? Have they used the kettle, or the shower, opened the fridge door, watched TV or listened to the radio? Monitoring all of these can enable analysis that can identify events where targeted interventions makes sense. Where a person has not spoken to someone for a while, a person can be sent round to have a chat. Where a meal hasn’t been had, someone who can prepare one for them. A fall? Send a first responder.
Such targeted interventions can ensure that a person gets the help that they require, when they require it. It can also help optimise use of the NHS’ scarce resources.
Even with a more concentrated care environment, such as a care home, such an approach can help in optimising the care a person receives. Heart rates, breathing, temperature can all be monitored. Even areas such as the condition of adult nappies could be monitored, sending alerts when these need changing. This again frees up the care staff to concentrate on the more human aspects of the job – talking and interacting with each person, rather than checking up on them.
But, all this needs some form of better standardisation around how data is held and transferred in such environments. Where near real time intervention is needed, any transformations of data can just slow things down. What KnowNow would like to see is more of an agreement around how datasets and APIs are created and managed – to make them more open, more available, more usable. This would be in the best interests of everyone involved – the IoT can only be effective where data is easily moved around and analysed.
I found the discussions with David very interesting – to me, KnowNow is one of just a few companies at the forefront of dealing with data in a manner that is suitable to the IoT. It is apparent that data formats and APIs will be a sticking point for a truly effective IoT – it is incumbent on all players in the market to ensure that data is easily and freely available from their devices.