The explosion of big data, and the general adoption of IoT
technology and cloud computing has called for faster, better
and more intuitive integration with geospatial analytics.
In 2017, Geospatial analytics market reached 38.65 billion . It
is estimated to grow at an annual rate of 18.2%, and by 2027,
the entire geo analytics market will reach 174.65 billion.
It’s no surprise that over 78% of C-level executives are
planning to invest in location data next year since the location
data is used for finding new markets and revenue sources,
delivering timely and tailored marketing campaigns and
detecting fraudulent activities in real-time.
For many enterprises going through the digital transformation,
it has been a real challenge to provide a unified solution to
both online transaction processing (OLTP) and online
analytical processing (OLAP) workloads simultaneously to
satisfy their core business requirements, and adding real-time
geospatial capability with seamless integration to existing and
future applications, with minimal redesign of data schema and
Why TigerGraph, a Native Parallel Graph Database for Geospatial Analysis?
Traditional Solutions Are Missing the Mark
Using traditional RDBMS for geospatial analytics has its
drawbacks: first of all, the data model is not designed for
complex location-based analytics questions in real time;
secondly, the geo analytics may rely on third-party index; in
addition, there are no easy SQL queries to answer these
complex location focused questions. Traditional geospatial
analytics applications built on relational databases were not
designed to address this challenge. Because of these
challenges, enterprises may not have taken geo analytics into
the core business logic previously, hence losing crucial
insights from the location data.
Graph databases, on the other hand, model our surrounding
environment as interconnected entities and explore the
relationship between people and things.
Geolocation data is
an intuitive addition to graph data modeling, which is naturally
designed to model the absolute and relative position of the
TigerGraph’s massively parallel processing or MPP
architecture empowers enterprises and governments to
expand their solutions, adding the location dimension(s) to
the use cases. While previous generation graph databases
struggle to provide analytics and insights that require more
than 3 hops or steps, TigerGraph’s Geospatial analytics can
easily traverse 10 or more steps to deliver new insights based
on the location data.
Consider a mobile offer or recommendation system that
factors in the customer movement data, where customers are
categorized into different mobility patterns based on their
movement throughout during weekdays and on weekends.
Popular commercial locations such as stores or restaurants
visited is also considered. Traditional marketing campaign
solutions lack the capability to combine customer
demographics as well as browsing, search and purchase
history with their mobility profile and deliver targeted offers
based on the real-time location of the customer.
TigerGraph based marketing campaign solution combines the
transactional data from sales and service channels containing
browsing, search and purchase history, with the location
history data indicating the most popular commercial locations
such as stores and restaurants for the customer. All of this
information is combined with the real-time location of the
customer and a tailored mobile offer (such as 30% off at GAP
store in downtown San Francisco or Tokyo) is computed for
real-time delivery to opt-in customers.
In order to detect these hidden relationships among location,
time, product, transaction and customer data, TigerGraph
executes a deep link query with 4 to 6+ hops going across
multiple connected entities to find the location-based product
or store recommendation. Traditional geospatial solutions
built on the relational databases may struggle to integrate
location data due to the rigid schema and require
computationally intensive database joins, rendering deep link
Geospatial applications require real-time location insights. Imagine a ride-hailing company which tries to identify vehicle supply and demand mismatch, to predict and resolve long passenger wait time and reduce vehicle idling time in a busy metropolitan area. Real-time geo insights from TIgerGraph can identify these issues and recommend actions to resolve them. A large amount of geodata is collected continuously. TigerGraph’s highly performant and scalable engine can analyze and detect the movement patterns of many people and vehicles at the same time and respond in real-time to the demands from millions of customers moving across a dense Metropolitan area such as New York, London, and Tokyo.
Geospatial Analytics for the Internet of
Geospatial analytics is particularly relevant for the Internet of
Things, as all the resources (sensors, switches, routers) in
the network are connected and there is a location tag
associated with the collected information. With TigerGraph,
you can process signals coming in from all of your IoT
sensors, actuators, switches and routers, map those based
on their locations and perform complex calculations in real-
time to make decisions. A great application of this is a faster
than real-time energy management system which combines
geotagged IoT data to balance supply and demand.