Best Practices Using Analytics and Data for Good: Disaster Response Case Study

Transcription

1 Best Practices Using Analytics and Data for Good: Disaster Response Case Study Brian Kelly Head of Community Stabilization Unit International Organization for Migration I-Sah Hsieh Global Manager, Int l Development SAS #AnalyticsX C o p y r ig ht 201 6, SAS In sti tute In c. Al l r ig hts r ese rve d.

2 AGENDA BEST PRACTICES DATA FOR GOOD Data4Good overview Role of Data Role of Analytics Case studies: Disaster Responder s analytic journey ACTwithSAS a crowdsourcing initiative

3

4

5 NEW FRONTIER BIG DATA GETS BIGGER!

6 FIRST WAVE MACHINES TALKING TO PEOPLE

7

8 SECOND WAVE MACHINES TALKING TO MACHINES, THEN TO PEOPLE

9 EXAMPLES INTERNET OF THINGS ANALYTICS Smart Grid Smart Home Smart Car Smart City Connected Devices Connected PMUs Connected lock Bluetooth dongle reading vehicle data Sensors embedded in road surface Data Insight Detect voltage instabilities on transmission network Detect when I leave Predict unsafe conditions & likely collision Detect weather, predict traffic, sense 911 vehicles Action Trigger corrective actions to maintain grid reliability Lock the door, set the alarm and adjust the HVAC Slow the vehicle, alert the driver Adjust speed limits and shift lane directions to optimize traffic flow

10 BIG DATA ALTERNATE DEFINITION Big Data = More data than what you currently use!

11 DATA 4 GOOD BEST PRACTICES Where to begin? Answer the hard questions OR Let the DATA tell the story Use Analytics Visualization, Data mining, Forecasting, Predictive, Optimization Use more & new data sources (non-traditional data sources, BIG data) Consider new technologies (mobile devices, Cloud computing, etc.)

12 ENGAGING MIGRATION & DISPLACEMENT DATA

13 INTERNATIONAL ORGANIZATION FOR MIGRATION (IOM) 1953

14 INTERNATIONAL ORGANIZATION FOR MIGRATION (IOM)

15 CASE STUDIES IOM S ANALYTIC JOURNEY Improving INSIGHT, HINDSIGHT, and FORESIGHT: Typhoon Haiyan (2013) Nepal Earthquake (2015) Mediterranean Migration Crisis (2015) Cyclone Winston (2016)

16 DATA DISPLACEMENT TRACKING MATRIX

17 VISUAL ANALYTICS WHICH SITES HAVE MOST CRITICAL NEEDS? FASTER INSIGHTS

18 VISUAL ANALYTICS MOST COMMON PHRASES IN HEALTH PROBLEMS FASTER INSIGHTS Elderly

19 VISUAL ANALYTICS HOW DO WE GET PEOPLE HOME? FASTER INSIGHTS

20 DEEPER HINDSIGHT LOOKING ACROSS 4 TYPHOONS IN PHILIPPINES

21 VISUAL ANALYTICS PREDICTIVE ANALYTICS GENDER BASED VIOLENCE IMPROVE FORESIGHT

22 VISUAL ANALYTICS PREDICTIVE ANALYTICS CAUSES OF DEATH (TEXT) IMPROVE FORESIGHT

23 BIG DATA SOCIAL MEDIA & GLOBAL TRADE DATA

24 CROWDSOURCING ACT WITH SAS - HUMANITARIAN ASSISTANCE When disaster strikes Time is critical Manpower is limited Analysis is often delayed External help is often unorganized

25 CROWDSOURCING ACT WITH SAS - HUMANITARIAN ASSISTANCE What if we could mobilize thousands of people to: Review data? Answer questions using leading analytics tools? Bring new questions, data, and ideas? SAS provides the cloud-based analytic platform Help on their own schedule No travel required (only internet)

26 CROWDSOURCING EXAMPLE: MISSING MIGRANTS PROGRAM In first half of 2016, migrant deaths increased 28% Why? What are the dangers migrants face by route? By region? By demographic? Small team (<3 people) Based in Berlin, Germany

27 CROWDSOURCING DEMO

28 #AnalyticsX C o p y r ig ht 201 6, SAS In sti tute In c. Al l r ig hts r ese rve d.