Big Data for Better Innovation

Post on 13-Feb-2017

219 views 3 download

Transcript of Big Data for Better Innovation

Big Data for Better Innovation

Professor Yike GuoDirector, Data Science Institute

Imperial College London

Service Availability Service Richness

Anywhere

Inte

ract

ion

bet

wee

n p

arti

cip

ant

s

Anything

Anytime

Service Accessibility

Anytime Anytime

AnywhereCloudCloud

WirelessWireless

Big Data

Digital landscape: towardsa digital service economy

Cloud: computing infrastructure

Big data : contents

Wireless : interaction platform

Digital landscape: structure

Technology_ Virtualisation, big data management(NoSQLDB), WebService and PAYG support

Impact_ Revolution in IT resource provision_ Unlimited, centralised computing capacity

Future trends_ Cloud computing: the main IT provision

mechanism_ Green data centre: equipped with power

computing resource_ In memory data management and analysis:

providing crucial support

Cloud Computing:cheaper and better

Technology_ High throughput mobile networks; new

mobile devices; new mobile Internetservices

Impact_ Revolution in social media and Internet

computing; sensor + wireless technologyenables real-time datafication; mobiledevices as the platform of interaction(human <-> services, data <-> physical)

Future trends_ 5G wireless network_ Context aware service provision_ Ambient intelligence_ Big data integration

Wireless technology:ubiquitous

Technology_ Pervasive sensing technology, big data analytics,

wireless communication, cloud computing

Impact_ On scientific research; real-time decision making;

building a valuable data asset

Future trends_ A new ecosystem: sensor fusion,

software agents and community sensing_ Building composed data products: interaction and

integration of quantification_ Real-time decision support services on the mobile

net

Big data: dataficatingeverything

Quantifying physical objects through measuring their features

How are innovation processeschanging?

How are innovation processeschanging?

I-o-T Cloud Mobile DataDistributed Digital Global

Anytime Anywhere Anything

I-o-T Cloud Mobile DataDistributed Digital Global

Anytime Anywhere Anything

[Parmar, Mackenzie, Cohn and Gann, Harvard Business Review, January/February 2014]

Business model innovationNew patterns of innovation: Using data to drive growth

[Parmar, Mackenzie, Cohn and Gann, Harvard Business Review, January/February 2014]

1. Augmenting products togenerate dataWhich data relates to yourproducts and their use?

2. Digitizing assetsWhich assets are wholly oressentially digital?

3. Combining data withinand across industriesHow might data becombined with that held byothers to create new value?

4. Trading dataHow could data bestructured and analysed toyield high-value insights?

5. Codifying a distinctiveservice capabilityDo you have distinctivecapabilities that others wouldvalue?

Data-driven innovationHow to use data to drive growth

Neuro-iBroadcasting

Augmenting media with viewing feedback

Technology_ Wearable sensors: capturing personal

physiological and behavioural information_ Cloud: data analysis

Impact_ Enabling real-time health monitoring and

behaviour characterisation_ The foundation: personalisation of products

and services

Future trends_ Ecosystem: personalised services_ Integrated data products: combined with

personal biological data for personalisedmedicine

_ Real-time decision support: mobile healthmonitoring

Digitizing life: Body Sensor Informatics

Body Sensor Informatics for Homecare ofmental diseases

Using body sensors toassess gait in progressivemultiple sclerosis patientsin their homeenvironments.

3-Axisaccelerometer

s

Continuous and dense measures in home environment Calibration with personalized clinical model

Access of MS Disability

Traffic flow, car emission data and weather conditionwill generate a dynamic map for the air quality of acity.

Digital City Exchange : Combining datawithin and across city sectors

Data storage platform forcommunity sensing:

– Collecting the sensor data in a wikiway: everyone can store any data ifit’s describable (by ontology)

– Data storage and managementinfrastructure with highperformance, low cost, and goodsecurity using the Big Dataarchitecture

– Universal Query Language (UQL) forretrieving data from varioussources, and acquiring data ondemand pollution

weatherhealthtraffic

energy

social network

WikiSensing Data Storage

UQL

Wikisensing: Combining data within andacross city sectors

Pay-by-Data Model: enable data tradingFour components:•data collection service•data pricing agreement•authentication service•interface to app and app marketplace

When an application is submitted to theapplication marketplace, the agreement ofdata usage is also submitted. Here is anexample of such agreement (json):{“app_id”: “ios.1154431”,“url”:“http://wikisensing.org/client/appexample”,“scope”:[“precise-location”],“frequency”: 21600000,“amount”: 10,“valid_time”: 684000000,“operation”: “read_only”}

Imperial College Data Science Institute:A Focal Point

DataScienceInstitute

STRATEGIC APPLICATIONSFACULTIES

Health, Wellbeing &Personalised Medicine

DiscoveryScience

SustainableDevelopment

Energy & Environmentof Future Cities

Faculty ofEngineering

Faculty ofMedicine

Faculty ofNatural Science

Imperial CollegeBusiness School

Codifying our multidisciplinary capacity forbetter science

Thank you

Data Science InstituteImperial College London

Professor Yike Guo