Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

39
Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK

Transcript of Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Page 1: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Using Real Time data to Understand and support Human Behaviour

Paul WatsonNewcastle University, UK

Page 2: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Real time Data Sources increasing

Sensors

Environmental

Medical

Location

CCTV ANPR

RFID

In Car

Page 3: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Real time Data Sources increasing

People (directly)

e-mail

twitter

text

Page 4: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Real time Data Sources increasing

Software as a Service

People (indirectly)

Page 5: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Challenge

How can we use real-time information to influence behaviour (for the good)?

Page 6: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Key Stages

How to model and reason about behaviour?

How to influence behaviour?

Intervention

Sensors

What data to collect?

Examples from Recent Work at Newcastle.....

Page 7: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Social Inclusion Through the Digital Economy (Hub)

Social Inclusion

Digital Technologies

Page 8: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Professor Phil BlytheVisalakshmi Suresh

Page 9: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.
Page 10: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

CENEX Electric Smart Car

Page 11: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

GPS &

Engine Management Systems

Real-time Vehicle Monitoring

Existing Traffic

Management Systems

Real-time Data Management

(at Ncl University)

Page 12: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

TRIP SUMMARY

1. Power Consumed (.0611KWh)

2. Average Temperature (4)

3. Break Pedal Pressed (996)

4. Distance (8.4 km)

5. Speed (34 km/h)

6. Carbon foot print (6.972 g)

TRIP SUMMARY

Page 13: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Battery Depletion

Integrate with Satnavto guide driver to charging station

Page 14: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Assisting Older Drivers

• Can data analysis determine long-term issues?– cognitive impairments

• Short-term problems– effects of drugs

• Suggest the need for assistive technologies

Page 15: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Influencing Driver Behaviour

How to model and reason about behaviour?

How to influence behaviour?

Intervention

Sensors

What data to collect?

Engine ManagementGPS

Identify patterns that require interventione.g. Distance Alert

Page 16: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Professor Patrick Olivier, Newcastle University

Can we use pervasive technologies to allow people to stay longer in their own homes?

Page 17: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Supporting people with dementia

• People in the early stages of dementia want to continue living at home.

• They have problems with:– episodic memory – executive control (planning,

sequencing, attentional control)

“Gets the kettle, fills it, switches it on…But then she’ll stand there, and I’ll say – ‘what are you doing now?’ and I’ve jogged her memory again…She gets the coffee jar, and she’ll take the top off the jar. And she’ll look at it and think ‘what am I doing with this off for?”

“Gets the kettle, fills it, switches it on…But then she’ll stand there, and I’ll say – ‘what are you doing now?’ and I’ve jogged her memory again…She gets the coffee jar, and she’ll take the top off the jar. And she’ll look at it and think ‘what am I doing with this off for?”

Page 18: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Ambient Kitchen

Page 19: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.
Page 20: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Activity Recognition Video

Page 21: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Influencing People with Dementia

How to model and reason about behaviour?

How to influence behaviour?

Intervention

Sensors

What data to collect?

RFID Accelerometers in utensilsFloor pressureVideo

Data Mining,Activity RecognitionPartially Ordered Markov Decision Processes

PromptingAlerts

Page 22: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

e-Science Central

• Software as a Service for e-Science

Page 23: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Software as a Service

Sharing can be integral part of application- in real time

System receives info from all users in real time• challenge is how to influence user behaviour

Real-timedataexchange

Sensors

Actuators

Page 24: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

e-Science CentralStore, Analyse, Automate, Share

e-ScienceCentrale-ScienceCentral

• Dynamic Resource Allocation• Pay-as-you-Go*

•Web based•Works anywhere•Web based•Works anywhere

•Controlled Sharing• Collaboration• Communities

•Controlled Sharing• Collaboration• Communities

Page 25: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.
Page 26: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.
Page 27: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.
Page 28: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.
Page 29: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.
Page 30: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Blogs and links

Page 31: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

e-Science Central – Social Networking

Page 32: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Provenance powers Dashboards & Collective Intelligence

Page 33: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Advise on best practise : “Most of your colleagues use workflow W to analyse the type of data you’ve just uploaded”Alert when interesting new data appearsPutting people in touch with experts

Influencing Behaviour of Scientists

How to model and reason about behaviour?

How to influence behaviour?

Intervention

Sensors

What data to collect?

Provenance (store, analyse, share)Social Connections

Graph theory, collective intelligence, provenance analysis

Page 34: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

3 Key Technical Challenges

Page 35: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Understanding and Influencing Behaviour

Real World Computer Model

Capture Model

Reason

Influence

Computer ScientistsDomain Researchers

Social Scientists

Page 36: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Personalisation• Influencing requires personalisation• There has been recent work on algorithms to

analyse vast amounts of data– e.g. collective intelligence, web analysis

• Assumptions behind most of this work:– no privacy issues– results not needed in real-time

• More focus needed on personalisation & timeliness– presenting useful information, in real-time, observing

privacy

Page 37: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Scalability

• Real-time data pattern matching & processing complex event processing?

• Historic Analysis cloud?

Page 38: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Real Time Historic

App App App

Consumers

Consumers & Generators

App

Sensors

Generators

App

Data WarehouseEvent Processing

Our General Architecture

Inform

Create ModelsCalibrate ModelsAggregations

Filter

Sensor events→ Application events

Page 39: Using Real Time data to Understand and support Human Behaviour Paul Watson Newcastle University, UK.

Summary

• Real time data increasing– Sensors– Software as a Service

• Key is to extract value from this data by understanding the real-world behaviour it represents

• Grand challenge is to use this information to influence behaviour (for the good)