“Where” and “when” are the typical answers from the Geographic Information Systems (GIS). But how do we make this data relevant when data flows become so large that no solution can deliver intelligent information for proper decision-making? What defines smart data?
With data processing tools, we are not only waiting for a thematic display, with zoom and filtering of the presentation but also to receive notifications of events adequate and circumstantial, for better decision-making. The intelligent society of smart cities demands a global vision, placed in the right context. In this structuring, the notions of “when” and “where” are particularly important for understanding the “why” and the “how”, human behaviour being entirely dependent on space and time.
Moore’s Law inapplicable to Big Data
For 20 years, the uses and functions of GIS (Geographic Information Systems) have largely evolved and expanded as tools evolved. This is only the beginning: it is impossible for an existing GIS system to remain in service without fundamental change in the next 10 years – and probably less.
We face such an explosion of amounts of information, backgrounds and very different formats, to be dealt with without transformation, despite increased almost consistent with Moore’s Law the power of computer processors and The advent of Cloud computing : sensors of IoT (Internet of Things), social networks, Open Data”, mobile terminals … the flow of information progresses too quickly.
Veracity and Value give meaning to the data
GIS, like other technologies that are transformed with Big Data, must respond to the 3vs: volume, variety and velocity. But to assimilate and transform these data almost in real-time, we must add 2vs complementary: veracity and value.
This means context and relationships between the data and their quality define their relevance and value in decision-making, compared to a mass of isolated data which is thus almost useless. The first step in giving them meaning is a reliable and coherent geospatial context to position an event where it actually occurs.
From Big Data to smart data
To create wisdom to have vision and to act intelligently from data, we need to develop an intelligent and scalable rule-and-filter system that will transform the Big Data into bytes of smart data:
Recovered data, though often derived from well-structured corporate or community systems, must be combined with other sources, such as Open Data, IoT, or social networks, whose structuring is not always the same.
Automated rule systems must be flexible enough to reveal and assimilate new data formats and models and make them workable, or even be able to complement existing data by developing logical reasoning.
Set a goal to determine the right filters
It is a matter of giving an objective to the data. It will determine the direction of the rules that will automatically sort the exploitable data. For example, in order to understand the evolution of the filling rate of the hotels of a city such as New York and to define the actions to increase this rate, the filters will have to integrate the prices of train and plane tickets to this destination, The good state of tourist areas in the city center, but ignore the renovation of a road tunnel linking the city to its suburbs.
Only intelligent data planning will meet this challenge and present synthetic dashboards, including reports, charts and maps, with different levels of filters transforming big data into KPI (Key Performance Indicator) to get a global view. Flexible and adequate rules and automation will provide the right context and adapt products in a short time to the needs of the market by a few adjustments.
Bring out healthy sources
Once the objectives have been established, filters can be created to clean the data, remove noise, supplement them and act on refined sets. But this is just the beginning. Beyond the rules, one has to be able to trust the data collected. Deceptive (truthful) or poor quality (value) data will not help decision-making, on the contrary. Some modern, scalable, intuitive, open standards-based rule engines can help sort, clean, add, add, and extract KPIs or event logs.
The process is more important than the result
What really matters is the process; it is the process that validates the results. The tools and methodologies used are primarily useful to give meaning to the data retrieved. The most modern GIS adopt a positive rules approach and adapt to any type of format. Data that does not conform to pre-established rules will be identified and automatically retrieved.