Location Intelligence Platforms are the next tech frontier (on the same scale as search or e-commerce) as mobile, wearable and Internet-of-things (IOT) technology have been prototyped, experimented with and deployed in real-world scenarios. These platforms will help to answer the following – What? When? Where? Who? – with respect to events in the real world.
As a first step, it is important to understand the evolution of these location-centric data platforms. The earliest location intelligence platform is/was a “map” – an atlas which over the past two decades has evolved into digital maps for navigation – in custom hand-held devices and now, mobile apps along with the wide-spread adoption of GPS in mobile handhelds. The other major location intelligence data source is/was satellite data, mainly used for remote sensing, land use analysis and mapping weather on a global scale. Maps provided a static view of the physical world at different levels of resolution, whereas satellites provided a slowly changing view of the world at a large scale; changes in weather, vegetation, topography etc. However, the whole field has had a new impetus with the availability and usage of GPS embedded mobile devices – in phones (used by people) and on transportation systems (auto systems, trucks, planes, ships etc.) at a global scale.
Given these individual real-time sources of location data, and possibly additional information available at a location via static sensors (even marine buoys stream weather information), it is possible to monitor and track what is happening, what has happened and possibly predict what may happen at every location globally, along all dimensions – latitude, longitude and altitude/depth (both above and below ground). If it is possible to harness these disparate information sources, one can envision that it is possible to know what a single individual or groups of individuals do on a regular basis in a city versus a village, what happens at Times Square in New York or Piccadilly Circus in London every hour of every day or what is bought on Bond Street versus your local market.
Location Intelligence is the gamut of models, algorithms, techniques and tools that assimilates all the relevant data sources, processes the data and provides requisite answers to such aforementioned questions. Location Intelligence Platforms abstract the above technologies and provide it as data-as-a-service.
A Location Intelligence Platform is powered by both static and dynamic sources of data. Static sources are primarily digital-map databases: cartographic, topographic and satellite data. Additional static sources such as real estate databases and public land-use databases are also coming online. Dynamic sources include IOT data feeds, social data feeds, mobile app data feeds and telco data feeds amongst others. Both types of sources include data in structured and unstructured data formats – images, videos and text. However, core technological issues may include:
● Harnessing the aforementioned sources and de-noise them to improve the base data quality
● Linking them to spatial and temporal reference frames at different levels of resolution
● Adding context to the raw data from different entities – people, vehicles, products, weather, pollution etc. – and enable semantic reasoning at higher levels of abstraction
● Build “models” to reason with both high-level and low-level data – infer phenomena such as modes of transport or folks visiting a particular retail location only on weekends
● Store and persist this data to make it amenable for querying about the past, the now and possibly the future to infer micro, macro, individual and group trends, and
● Finally, handle the scale of the incoming data in a reasonable, cost-effective manner.
Each of these technological needs is pushing the technology envelope in different areas:
a) Spatio-temporal modelling, databases and warehouses
b) Handling raw data and processing these event streams at scale
c) Monitoring and detecting trends – pushing the frontiers of statistical modeling, artificial intelligence and machine learning
d) Algorithms to process streaming data to characterize the underlying phenomena
e) Mobility modelling of entities and other locational phenomena.
Building a useful location intelligence platform requires a highly inter-disciplinary approach and there are any number of interesting problems to work on, both from a systems and modeling perspective. Though there are a number of academic research efforts, the explosion of commercial applications based on location has suddenly made this area a mainstream activity.
Many recent technological innovations could benefit from the next generation of Location Intelligence Platforms including applications such as Uber-like resource utilization systems for transportation, logistics and other services, large-scale crowd-sourcing for traffic and event management, news gathering, urban planning, drone-based applications, automated delivery systems, supply chain management, consumer applications for services and goods, including travel, hyper-local services, advertising and many other applications yet-to-be envisioned.
Considering that our lives are played out and embedded in the physical world, the development of the next generation of Location Intelligence Platforms to understand the same is a worthwhile technological endeavour.
