defining data clusters for transportation benefits
TRANSCRIPT
BIG Data Transformation for
Transportation Outcomes
DAVE VERMA – HTTP://LINKEDIN.COM/IN/DAVIDVERMA
The Opportunity…..
Government Agencies currently collect massive amounts of Data from a
variety of Transportation Systems using ITS equipment.
The sheer volume of data ( Big Data) is growing exponentially and is so
diverse that the ability to transform this into meaningful business
intelligence that adds value is compromised or nearly impossible under
current conditions
A new method and paradigm for both handling this data and applying it is
required – this methodology is needed to:
1) Create value from what has been to date a vast underutilized resource
2) Remove traditional thinking around what and how data is seen, accessed
used and more importantly rationalized
3) Enable the next generation of distributed heuristic logic processing –
Artificial Intelligence - to effect real world changes Dave Verma – http://linkedin.com/in/Davidverma
BIG Data = Bigger Decisions = Bigger Risks
Data Lifecycle – What is the Expiry Date?
Data Retention –Base Data or Processed Datasets?
Cluster Parity – When is the Data in a Cluster Correct?
Access – Whose data is it, before its processed and then after?
Point in Time access – data may be sensitive today but not tomorrow,whom decides?
How do you ensure resilience, do you protect source data or meta views?
Human Factors – what are the roles of the Actors in the process and how
far do you choose to automate or hand over to AI?
Dave Verma – http://linkedin.com/in/Davidverma
Clusters = the future of Big Data
Big Data can not be handled with traditional methods.
To make sense of Big Data it needs to be rationalized and normalized into meaningful Clustered Packages which allow for stateful use.
Vertical Clustered Data Packages create a manageable identifiable subset of Big Data
Clustering requires a number of key elements for it to work:
1) Concept of Cluster – what does the cluster collect and what does it do with what has been collected, how is this going to be tested and what are the outputs from it?
2) Once the COC and Validation process have been defined then the more detailed work on the Cluster Functional and non Functional Requirements and Preliminary Designs can be undertaken
3) Following the process above the detailed cluster dataset can be designed and relevant heuristics developed to interrogate and functionally operate the cluster. Automated test procedures can be run to ensure all outcomes are within expected ( designed) parameters.
Dave Verma – http://linkedin.com/in/Davidverma
The Roadmap to Effect Change:
Targeted Heuristics
Smart Clustering
Dynamic Private Clouds
Specific AI Applications
Big Data
• Asset Management
• Incident Response
• Emergency CIMS
• Dynamic UTC
Controls
• Info Channels
• Enforcement
• Security Systems
• Planning
Real World
Applications
Dave Verma – http://linkedin.com/in/Davidverma
Clusters Lead to Business Outcomes:
Dave Verma – http://linkedin.com/in/Davidverma
Transport Data Sets
Decision Support (AI)
Asset Management
Transport Clusters – The New ITS Paradigm -
Drive Content & Services not Devices
Dave Verma – http://linkedin.com/in/Davidverma
• Commodity ITS equipment produces high volumes of data
• The Clustered Data can create compelling content
• Combined data from different systems creates valuable
Content for Transport Agencies, public, media and advertisers
The New ITS Paradigm – Its all About the
Back Office not the roadside
Dave Verma – http://linkedin.com/in/Davidverma
• Clustering requires different skills from Traditional ITS
• Back office architecture, definitions and operations are the new
frontier for ITS – the traditional business of signs, traffic signals and
lane controls is obsolete and will be replaced with vehicle based
systems.
• The role of communications networks and roadside devices will
become commoditized and no longer a specialist area.
•First commercial Autonomic Parking Vehicles
•V2V becomes a standard
2015
•Traffic Management is run by AI and uses V2V and V2C to operate
•Driverless Vehicles become widespread
2020
•Drivers are now obsolete all driving is now AI based
•Focus is entirely on ensuring reliability of road assets and infrastructure
2030
The Big Data Timeframe? 15 Years of
Extreme Change is Imminent…..
• Journey Time Reliability• Asset Management• Smarter Content Delivery• Big Data becomes priority
• V2V reduces human factors• Asset Management Needed• AI starts to use Clustered Data• Big Data becomes BAU
• AI removes human factors• Asset Management Critical• Environmental Concerns Critical• Road side ITS no longer required
Significant advances in Processor & Power technologies increase AI capabilities
The End
Thank You