The Internet of Things has been experiencing the full on hype phase as the ceaseless barrage of statistics as to the numbers of things that will be connected. Fortunately this is fading out to be replaced with meaningful experiences on where, and how, the business value is to be found. The Analytics of Things, or AoT, is a new, and welcome part of this shift delivered from some well-respected data analytic technology companies.
The introduction of the term AoT draws attention to the new data and its value being generated both by IoT devices, and Smart Services. As with all data there is a requirement to analysis, but in this case it will introduce new types of Analytics. AoT also refers to the use of analytics in making connected devices smart and able to take intelligent action. For more details see Deloittes; AoT – a short take on Analytics
Previous blogs in this series have dwelt on the new and innovative ability to use ‘real-time’ data to be able react in an optimized manner not previously possible. The introduction of Smart Services as the mechanism for this, using Complex Event Processing is a significant change enabled by IoT sourced data. The most recent blog illustrated the application of these capabilities in a Smart City Technology and Business conceptual architecture, and included the need for the advanced Analytics of AoT.
IoT is still in the development stage of understanding for many people so adding the even less understood topic of AoT seems almost counter productive. If anything seems familiar it will be the Analytic part though this is wrongly assumed to be part of Big Data Analytics. AoT is concerned with wholly different data which in turn has the capability to reveal very different outcomes from existing reports.
In the blog on Smart City Architecture a simple Smart Service use case ‘Find me Parking’, demonstrated interactions between Smart Services by adding a further Smart Service providing route planning with traffic jam avoidance. Obviously these services are aimed at Citizens, but there are substantial benefits to the City Administration from analyzing the radically different data that Smart Services produce. Three simple examples of the data available are;
- Query inputs from citizens providing their starting geo-location as well as their intended destination geo-location
- The true total demand for Parking including requests that were unfulfilled, as well as the amount of time a car was on the streets waiting for a parking bay to become available.
- The routes used and the changes introduced by this traffic, as well as journeys abandoned due to traffic or non availability of a parking bay.
It would be difficult, and expensive, to obtain any of this data by other means such as personal surveys, or checkpoints. But the real point is that each of these three data outputs is generated within the context of an event, or citizen activity; and shows both failed, and successful, activity event chains and outcomes.
AoT explores new sources of data from Smart Service requests, and outcomes, adding entirely new, and original, data sets that have not been available from traditional sources. IoT enabled Smart Services bring the capability to apply entirely new high value analytics.
There is also data available from direct-coupled IoT sensors, many of these sources are now new, and examples include; sensing Buses passing route points to check if they are on time; various traffic flow sensing schemes; traffic light maintenance and other directly sensed conditions. All produce data inputs to AoT, but often the individual sensors have too low values for AoT advanced Analytics. While valuable these inputs remain more suited to validate existing planning and scheduling applications, and reporting analytics.
AoT is not yet a fully developed set of technologies products and methods, but the possibilities are becoming increasingly recognized as IoT enabled Smart Services are becoming more widespread. Interesting articles have started to appear outlining the importance of the topic, with the crucial association to ‘Smart’; the Wall Street Journal Blog The Analytics of Things and Deloittes Short Take on Analytics are examples.
Some companies that are well respected in the ‘data’ sector are active proponents of AoT; including IBM, and Teradata, each having developed specific approaches and products for this wholly new branch of Analytics. Other technology companies together with Startups are active, and there is a strong link between AoT and AI, or Artificial Intelligence.
The IBM ‘Watson’ initiative, is an example of this broader approach being defined as a cognitive reckoning engine thus fusing the broader implications of AI with AoT. However the IBM approach certainly does contain the mix of people, unstructured data and IoT based smart inputs that are the basis for AoT. IBM’s acquisition of the Weather Channel is an interesting example of simple sensor data being turned into a Smart Service of Weather Forecasting. The forecast data is then consumed in further Analytical analysis in combination with other Smart Services.
IBM, in common with other providers of AoT capabilities sees this new generation of unique outcomes being used to construct and refine the rules for Smart Services Complex Event Processing engines. This continuous refinement made possible by the interactive relationship between AoT, IoT and Smart Services is claimed to lead in the direction of Cognitive Reckoning and AI.
The earlier example of the Smart Parking and Smart Routing Services feeding new data to AoT illustrated just how this provides a very different view of parking and traffic activities. Using these results to adjustments City Planning and traffic models will lead to new insights being made available to both services to further refine the optimization of their outcomes. AoT adds significant new value, but will also need significant new skills.
Teradata have recently introduced a new business unit specializing in AoT deployments, this builds on the release at the beginning of 2014 of Teradata Aster-GR to add Graph capabilities to the flagship Teradata Aster Discovery Platform 6. Shifting to managing and exploring data by means of Graphs is an important principle underpinning much of the use of IoT data, and therefore AoT, see blog From Relational Databases to Graph.
Teradata used their online magazine to publish a description of the Aster-GR capability and how it integrates with other data in different formats in an article entitled ‘Hands-On’. Teradata has a more technically detailed paper entitled SAS/Access – interface to Teradata produced in conjunction with SAS, (a well respected name in advanced Analytics) on how to integrate solutions.
If you work in Analytics and figure that Big Data, and Data Lakes, are as far as the technology has advanced then its time to take a fresh look at the topic. Google AoT, or Analysis of Things, to find a range of content on the new fresh capabilities that it introduces. As Smart Services, based on IoT, become established as the real definition of the Digital Service economy then AoT moves to the forefront of delivering business and competitive value.
An important part of the research is to carefully check out how AoT will function in association with an Enterprise Data Warehouse; the good news is that this may be a whole lot simpler than expected. As an example Teradata accepts direct IoT data inputs for processing and aggregation into the existing data.