IoT: Ready, willing and able for the public sector?

SAS argues that the possibilities for IoT's use in the public sector are endless


By SAS

13 Dec 2018

The commercial world has been experimenting with the Internet of Things (IoT) for many years. IoT’s early days of development focused on transport and manufacturing and evolved to smart in-store retail customer experiences where sensors monitor buyer behavior and serve up offers to drive sales and consumer satisfaction. Whatever the industry, it seems there is always a viable and valuable use case waiting in the wings. That is in part down to the seemingly infinite learning possibilities that the analysis of IoT data opens up. Consider this simple statistic – a single train can now emit more than one billion data points per second on each journey – it’s clear that the learning possibilities from IoT appear almost limitless. Is the same true for the public sector, and if so, what is the one single success factor all organisations will need to embrace?    

IoT and you – a match made in heaven?

I think the answer to that is yes. Especially when you remember that the value of IoT data lies in its connectivity to devices that can generate huge volumes of data about the environment around them. It makes sense therefore that public sector organisations of all kinds should want to turn that data into insight and value. Take the simple example of DEFRA. Imagine how much new information the department could gather by deploying sensors in sensitive coastal ecosystems under mid-term threat of erosion. How about in marginal farmland environments to learn how to boost productivity? Or even in rural communities to assess patterns of transport and local service use in order to develop strategies that promote cohesion and prevent the growing health concern of isolation and loneliness. 

Suddenly, it’s very easy to see how IoT can make a real impact on our nation for the better. More intriguingly, it can also help to improve individual citizen outcomes ‘in the moment’ as well as delivering new learnings about the kinds of services citizens and organisations really require from government. Not to mention the far richer insights it will provide in order to help government departments better balance their own needs to do more with less, while improving efficiency and outcomes. Something that is especially prevalent in the hard-pressed NHS.    

Time to change your thinking on analytics

It is precisely IoT’s ‘in the moment’ potential to derive insights, impact interactions and improve outcomes that I believe offers the public sector most value. In the first instance, the DEFRA scenario, data would likely be transmitted back to a central repository and prepared for later use. For example, no insights will be derived, or decisions made at the time of a storm being monitored on a certain section of coastline. Instead, analysis will be undertaken after the event, perhaps to detect trends and effects on a cliff’s stability.  

However, let’s take the example of energy consumption in a city by publicly provisioned services, such as street lighting. Monitoring is continuous and contextual – synced with traffic flows (vehicles and pedestrians) as well as air quality. Systems will be able to interact with devices at control centres, managing the street lighting, switching it on and off according to demand, daylight conditions, time of year and potentially switching power suppliers depending on the cost and supply of energy available. 

In this scenario, decisions are made in real time, based on the data being streamed from multiple, interacting sensors. Because decisions and outcomes need to be made ‘in the moment’, data cannot be transmitted to a storage facility, prepared, and analysed later. The lag time would be too great. A new, real-time routine is required.

The critical success factor: A new analytics architecture for ‘in the moment’ insights

It’s important to note that with IoT data, not all of it will be relevant and some of it might not be gathered to the quality you would ideally like. What you will need is the ability to clean data in motion using pre-designed data quality routines and text processing so that it is ready to use. You’ll also require streaming analytics, probably combining open source and proprietary machine learning with high frequency types. These will allow you to connect, decipher, clean and fully understand your streaming data as it flows no matter the volume or types of data (text, voice, video and more). Arguably, most importantly, you will also need high performance processing capabilities, so that you can analyse data as it travels directly from IoT monitoring systems to your organisation. This could incorporate millions of events per second. Critically, as insights are needed ‘in the moment’, lag time in gaining analytical outcomes is not an option. 

Automation through artificial intelligence is the key

Because of the high speeds and high frequencies of this kind of analytical environment, you’ll require significant levels of automation and the ability to learn from the data that is live streaming. Finally, all this data should be filtered, normalised, categorised, aggregated, standardised and cleaned before it comes to rest and is stored. Why? Who in the public sector has the budget to spend on staff capable of sorting through vast data lakes, after the fact, not to mention the available hours to undertake such a task. 

Who can help you?

Finding a partner that can help you to acquire all this rich functionality, with the requisite governance frameworks in place is tough. I do note that SAS is capable of delivering the kind of event stream processes solution that could support your IoT ambitions. In fact, it is this AI-driven analytical platform complete with data management capabilities that is critical in transforming the wealth of IOT data into valuable outcomes whether for citizens, the organisations you serve, or your department’s strategic agenda. Frankly, we all need to de-risk our forays into advanced analytics and partnering with a renowned leader, is one brilliant way to do exactly that.

For more information, please click here 

Read the most recent articles written by SAS - Can ‘data driven’ be learned, or is it ‘in the DNA’?

Share this page