• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Taking the Global Pulse - Photo Book
 

Taking the Global Pulse - Photo Book

on

  • 915 views

In November 2011 at the United Nations Headquarters in New York, the General Assembly was briefed on the progress of Global Pulse, a UN innovation initiative harnessing real-time data and new ...

In November 2011 at the United Nations Headquarters in New York, the General Assembly was briefed on the progress of Global Pulse, a UN innovation initiative harnessing real-time data and new technologies ("Big Data") to protect vulnerable populations. This book is a record of that event and presents 1) emerging techniques in information technology that make Global Pulse possible, 2) preparatory work in Indonesia and Uganda for launch of the first Pulse Labs, 3) the results of five research projects to test concepts underlying the application of real-time data and global development and 4) the roadmap for UN Global Pulse

Statistics

Views

Total Views
915
Views on SlideShare
852
Embed Views
63

Actions

Likes
0
Downloads
12
Comments
0

1 Embed 63

http://postscapes.com 63

Accessibility

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Taking the Global Pulse - Photo Book Taking the Global Pulse - Photo Book Presentation Transcript

    • TAKING THEGLOBAL PULSEUsing New Data to Understand Emerging Vulnerability in Real-Time 1
    • In November 2011 at the United Nations Headquarters in New York, the General Assembly was briefed on the progress of Global Pulse, a UN innovation initiative harnessing real-time data and new technolo- gies to protect vulnerable populations. This book is a record of that event and presents 1) emerging tech- niques in information technology that make Global Pulse possible, 2) preparatory work in Indonesia and Uganda for launch of the first Pulse Labs, 3) the results of five research projects to test concepts underly- ing the application of real-time data and global development and 4) the roadmap for UN Global Pulse.1
    • The recent waves of global shocks – food, fuel, and financial – haverevealed a wide gap between the onset of a global crisis and the avail-ability of actionable information that allows leaders to make decisionsin time to minimize the effects of these crisis.Knowing the “how, where and when” of a crisis while it is still un-folding—particularly in regards to how people are coping with its ef-fects—informs efforts to increase resilience to global shocks and pro-tect hard-won development gains.Today, with the explosion of information and communication technol-ogies, particularly mobile phone-based services, communities world-wide are generating real-time data in ever-increasing volumes. Thesenew “data trails,” combined with our traditional data sources, holdtremendous promise for helping us listen for the early signs of socialand economic stress. 2
    • H.E. Mr.BAN KI-MOONSecretary-General of the United NationsI n 2009, at the height of the global economic crisis, it was clear that we OUR INABILITY The private sector is analyzing this new data to understand its customers in real- were seeing something new: impacts TO UNDERSTAND time. The United Nations must do the of the crisis were flowing across bor- THE IMPACTS OF A same for its constituents: people around ders at unprecedented velocity. CRISIS WHILE THERE the world who are losing jobs, getting sick and having difficulty feeding themselvesIn today’s volatile and interconnected IS STILL TIME TO and their families. Much of this data con-world, when crises emerge in one part of ADJUST OUR POLICIES tains signals that are relevant to develop-the world, they have the potential to rever-berate quickly around the globe and inflict AND PROGRAMMES ment. We must use it to tell us what is happening, while it is happening.immediate suffering on the poorest and THREATENS TO REVERSEmost vulnerable populations. It is as if so- YEARS OF HARD-WON Today you will hear about the excitingcioeconomic crises can now move almostwith the speed of natural disasters. DEVELOPMENT GAINS. work that the Global Pulse team has been doing in analyzing new data and building new technologies. You will learn about the The irony is that we are actually swimmingYet at a time when our need for policy team’s roadmap for the year ahead, as in an ocean of real-time information. Theagility has never been greater, our tradi- they begin implementing Global Pulse at explosion in access to mobile phones andtional 20th-century tools for tracking in- the country level in Uganda and Indone- digital services means that people every-ternational development simply cannot sia. where are contributing vast amounts ofkeep up. Too often, by the time we have information to the global knowledge ware-hard evidence of what is happening at house. Moreover, they are doing so forthe household level, the harm has already free, just by communicating, buying andbeen done. selling goods and going about their daily lives.3
    • The idea behind this initiative is simple:once we know what signals to listen for, THE TIME HAS COME FORwe will be able to “take the pulse” of vul- US TO BRING THE WORKnerable communities. This rapid feedback OF THE UNITED NATIONSwill help us understand where people andcommunities are in trouble, how they are FULLY INTO THE DIGITALcoping with global shocks, and how to re- AGE.spond while there is still time to preventharm. 4
    • Mr.PETER HIRSHBERGAnnenberg Center on Communication, Leadership & PolicyDr.ANDREAS WEIGENDStanford UniversityPETER HIRSHBERG:It is a hallmark of today’s digital world that Over a billion times a day when we publishmankind generates vast amounts of data. to a blog or social network, we signal ourEverywhere, all the time. opinions, our sentiments, and send clues that social scientists couldn’t imagine aThese are signals, which have the poten- decade ago.tial for unprecedented insight into our be-havior, economics, our health and wellbe- Also today, there will be more than 300ing. million credit card transactions — each is an economic signal.On this day, there will be as many search-es on the Internet as people on earth. Andeach of these creates a signal of our inten-tion and interest.5
    • THE MOST PROLIFIC We humans will send about 10 billion text messages today alone.HUMAN SIGNALING Every mobile phone sends a location signal which helps paint a picture about howDEVICE IS PROBABLY people move, how cities grow, about economic activity, the quality of life.THE MOBILE PHONE. In 2011 we will generate more data than all of mankind has since the beginning of his-THERE ARE 5 BILLION tory.ON THE PLANET, AND 4BILLION OF THOSE ARE IN Properly analyzed, these signals can help detect an economic crisis in its early stages, or the outset of an epidemic, before it becomes an epidemic.DEVELOPING NATIONS. That is why mankind’s data is a core global resource, which is at once inexhaustible and pervasive. 6
    • ANDREAS WEIGEND:L et’s talk a little about the data that is generated by these signals. You might say, that data is the new oil. Like oil it needs to be refined tocreate value. Like oil, an ecosystem hasgrown up around it. But unlike oil, datadoes not get used up. We don’t have toworry that we will run out of data! We can distinguish two kinds of data: so- cial data – data which people share on so- cial networks and online – and observed– or transactional – data.An example of social data would be peo-ple’s geo-location, or when people sharehow they feel about something. It has be-come trivially easy for nearly anyone in theworld to express anything that they want,and to distribute it to their friends, onFacebook for example, or publish it to theworld, on Twitter or a blog.The other kind of data is observed data.That is, transactional data, objective data.Examples range from the level of thefloods in Thailand today, to credit cardtransactions, international money flowsor the records kept by the phone calls wemake.7
    • PETER: Speaking of phone calls, here iswhat massive, simultaneous human be-havior looks like. These are visualizations WHEN YOU MARRYof anonymized telecommunications data MOBILE PHONE ANDover a city map. LOCATION DATA WITHThe map shows Rome in the year 2006, CITIES, YOU GET A REALduring the World Cup finals between Italy TIME SENSE FOR HUMANand France. The SENSEable City Lab atMIT, in partnership with Telecom Italia, BEHAVIOR AND HOW IThave anonymously mapped text messages CHANGES OVER TIME.and calls from Rome that day during thetournament. It was the first time we wereable to watch human behavior at scale, inreal-time, through phone call data.As the game ends, and as the first 15minute overtime period begins, the callsbegin to spike. Then during the second15 minute overtime, when Zidane gets thered card, Italy goes crazy - and we canactually see the volume of calls and mes-sages spike across the city! Real Time Rome, SENSEable City Lab, MIT The peaks and valleys in these maps of Rome depict the real-time telephone and text communication across the city during the 2006 World Cup competition. These telecommunication spikes are also indica- tors of human emotion and excitement. 8
    • WE CALL THESE SIGNALS “DIGITAL EXHAUST.”THEY ARE A BYPRODUCT OF OUR LIVES.ANDREAS: The marginal cost of creating the time scales of reporting data. In otherthis data is zero, and during the last ten words, the timescales of observing datayears we have moved from a data poor en- were much slower than the timescales ofvironment to a data rich environment. the underlying process itself.This means we’ve created an ecosystem We are no longer limited by the timescalewhere people can create value from this of observing reported data. We now candata. Why do we care? Well, it’s about see the world in real time.making decisions. Let me give you a verysimple example: Traffic. We now have real The important thing about this is that ittime traffic data. With this, we can deter- allows for experimentation.mine know how long it takes to get homefrom here, allowing you to make better Last century was the century of the physi-decisions and cities to be run more effec- cal sciences, where physicists did experi-tively. ments and measured the outcome. This century is the century of social sciences.Let’s go back now, and look at the con- Where we can run experiments in the realcept of ‘real time’. In the past, many deci- world, and where we have the instrumentssions were made complicated because of that can measure the outcome in real time.9
    • COMMUNICATION USAGE PATTERNSCAN REVEAL SOCIAL AND TRADEBEHAVIOR IN REAL TIME AT ALMOSTNO MARGINAL COST.PETER: Here is an example of what that not London or a global business hub.type of real-time measurement can reveal When the cost of a call to the Dominicanabout New York. Republic for an hour is less than the cost of a subway trip, the economics changeResearchers from SENSEable City Lab, the way people raise their kids; and inter-MIT mapped all the telephone and in- nationally change behavior.ternet connections from New York to therest of the world. The data paints a vivid NYC TOP CALL DESTINATIONS SO OFTEN, WE ASSUMEpicture of global connections, of a global by minutespulse. DOMINICAN REPUBLIC 11.7% THAT TECHNOLOGY ISBut what is even more interesting is if we MEXICO 9.1% FOR WEALTHY NATIONS. UNITED KINGDOM 7.5%go underneath the surface of the data, and CANADA 7.0% HOWEVER, IF YOUstart looking at the stories the information GUATEMALA 5.5% LOOK AT WHERE IT CANpaints about commerce flow, about emi-gration and about how families maintain ECUADOR 5.3% MAKE THE BIGGESTrelationships. JAMAICA 3.3% DIFFERENCE, AND INDIA 3.0% WHERE ITS GROWING GERMANY 2.5%It turns out the number one place NewYorkers call is the Dominican Republic, PHILIPPINES 2.4% THE FASTEST, ITS IN DEVELOPING COUNTRIES. 10
    • Effluent cellphone data can save lives PETER: The United Nations has already cited Africa as having theand prevent data blindness at a time greatest rate of mobile phone growth in the world and it’s alsowhen it’s most critical. where some of the most exciting work is going on to understand these signals, and shed light on economic development.In the aftermath of the Haitian earth-quake it was an analysis of cell phone Over the past several years, researcher Dr. Nathan Eagle, and hisdata that showed where the popula- associates from Harvard and MIT, have analyzed millions of phonetion had dispersed to; vital informa- records - all anonymously - in Kenya and Rwanda, and comparedtion to aid workers. them to survey data.And when there was a cholera out- They’ve learned that the pattern with which a person purchasesbreak some months later, these tech- mobile phone time - or “tops up” their credits - and in which de-niques were so well rehearsed, that nominations has proven to be a remarkably accurate surrogate forwithin 12 hours, researchers from the economic health in a population. They found that people who topKarolinska Institute in Sweden, knew up in lots of small increments are generally less well-off economi-where in the country people in that cally than those who spend a similar amount all at once.area were dispersing to and couldidentify which areas might get an out- By gathering information across a country it’s possible to see vil-break next. lage by village, city by city, week by week, the economic well being of a population. Which means that by looking at digital exhaust, an economist in the DRC, or Indonesia, or Ghana, can measure the effect of an intervention. Or work with a policy maker to recom- mend one.11
    • 1 Higher economic1 1 10 wellbeing1 1 10 1 11 1 1 Lower economic wellbeing 1 1 1 1 1 1 1 1 1 1 Can mobile phone usage patterns can reveal the eco- mobile phone time and their economic well being. If nomic health of a population in real time? two people purchase the same amount of time in a month, the person who makes many smaller transac- The way in which a population uses mobile phones tions generally faces greater cash-flow issues and is reveals many patterns of human behavior across less well off than someone who can afford a single time and location: how we move about, how often we larger transaction. Because this data is available in communicate, how we pay for service and when. Dr. real time it can act as a powerful complement to tra- Nathan Eagle and his associates found a strong cor- ditional means of economic reporting. relation between how a person tops-up (purchases) 12
    • DIGITAL DATA CAN BE REMARKABLYPOWERFUL IN DEVELOPING COUNTRIES, ANDIN TIMES OF CRISIS.PETER: Let’s consider Haiti again. After the This happens not only in emergency situ- ANDREAS: Indeed, community is impor-earthquake, it was data from hundreds of ations. Today, anyone can contribute to tant, and sharing is important. And inindividuals that literally rebuilt the map of the knowledge of the world by fixing or an- fact, all the data that people share canthat country. Through community mapping notating something digitally. This kind of also create an extremely powerful detec-and mobile phone data, primary roads and bottom-up citizen energy leads to the type tion system.then eventually smaller residential streets of innovation we see in San Francisco, orand camps of displaced people during the Singapore, or even in the slums of Nai- Health officials often need to make deci-emergency were mapped on Open Street robi. In the slum of Kibera, it was citizens sions under pressure and with minimal ac-Maps. This was a tremendous example of that created the first ever map of their own cess to data. Let us take a recent examplepeople sharing data, and working together. community, and then started adding infor- from Germany, where there was an e-coli mation to it; valuable information that can outbreak a couple of months ago. help organizations and officials make bet- ter decisions. And this act of engagement, built a greater community. In Kibera, citizens use open source technology technology to create the most detailed map ever of this dense and im-poverished neighborhood. Then, the mapbecomes basis of a system to collect news, community and crime information.13
    • June 2011: E-coli Outbreak in GermanyEpidemicIQ, a San Francisco startup, produced a visualization What you see is that:that shows how data from both news sources and from thousandsof blogs revealed the outbreak, in real-time. 1) the community data was more timely, andThe visualization reveals – with relatively high granularity – that 2) the community data is of a much higher granularity than thethe outbreaks were clustered in northern Germany. At the bot- official information that came with delays and low reporting gran-tom of the graph, the red curve shows the number of outbreaks ularity.as reported by official health agencies. The green curve depictsEpidemicIQ outbreak estimates gathered from blogs and otheronline news sources. 14
    • SOMETIMES WE GET JADED BY THE PACE OF TECHNOLOGICAL DEVELOPMENT, BUT THESE SIGNALS, AND THE DATA THEY PRODUCE CONSTITUTE ONE OF THE FASTEST GROWING PHENOMENA ON EARTH TODAY.ANDREAS: PETER: So great is the production of this As an observer from the private sector and data, that many are calling it a revolution, from Silicon Valley where we often imag-A word about privacy: we have in- equal in import to the industrial revolu- ine what our technologies might do, this isvented something extremely power- tion. a moment of great hope and expectationful. And as with anything powerful it as we watch the United Nations share withis not black or white. It is our respon- Which is why work that the United Na- its many member states these new capa-sibility to communicate trade-offs so tions is doing in this area is so timely and bilities for all mankind.we can use these tools responsibly important. You’re bringing together twoand preserve people’s rights. cultures: the tools and systems to capture and analyze these signals at scale, and the human insights, the response mecha- nisms and government processes that will lead to better decisions. It is not an understatement to say that this is the century, and this is the moment when connection happened on earth! And the UN is providing the leadership in learning how to apply these in crisis iden- tification, to development and to health.15
    • IN 1995, THERE WERE ABOUT 50 MILLIONWEB USERS, THINK OF THAT AS THE SIZEOF THE MOON. TODAY, WE JUST PASSEDTHROUGH 4.3 BILLION WEB ADDRESSES.COMPARED TO THE MOON, THAT IS ABOUTTHE SIZE OF THE EARTH. AND IN 15 YEARSTHERE ARE FORECAST TO BE APPROXIMATELYA TRILLION DEVICES AND OBJECTS, ALLEMITTING SIGNALS, TELLING STORIES, ANDHELPING US ENGINEER A BETTER WORLD. THISTRILLION NODE “INTERNET OF THINGS” WILLDWARF TODAY’S INTERNET. BY COMPARISON,THAT’S THE SIZE OF JUPITER. 16
    • Mr.ROBERT KIRKPATRICKDirector, Global PulseG lobal Pulse envisions a future in Soon they begin calling agricultural hot Yet these digital signals are hard to ignore. which governments harness the lines in increased numbers, looking to They suggest that something is happen- power of real-time data to make sell livestock at a time of year when no ing to a vulnerable community. Something development decisions. Before we one sells livestock. Later still, they begin we need to know more about. At the verytell you how Global Pulse is working to receiving mobile money transfers from least, wherever these signals appear, isdeliver on this vision, let me first set the subscribers in urban areas. Eventually, where you need to conduct your next as-stage with a scenario that informs our ef- phones purchased in the rural area start sessment, to get the hard evidence youforts at Global Pulse. showing up in a nearby city. need for action.Imagine that it is 2009, at the height of Any one of these digital signals, observed The data we need is out there, and Globalthe global economic crisis. Food and fuel in isolation, might mean nothing at all. Pulse is actively working to learn how weprices have risen sharply this year, and Yet together, they tell an unfolding tale can use it for development. Even the datarains are late in coming. of a population struggling to make ends that is publicly available has tremendous meet, selling their assets, reaching out to potential to help us take the pulse of vul-How might the crisis unfold in the real- friends and family for help, and finally, nerable communities.time data collected by a mobile phone migrating for work.company in sub-Saharan Africa? We have analyzed public data on blogs, on Remember that at this point, we are look- social networks, and messaging servicesSuddenly, in a rural part of the country, ing at anomalies in collective behavior. such as Twitter. There are real signals inmobile phone subscribers who usually put We still do not know what is causing the there. We have found that some of what$7 credit on their phones every month changes. Have the crops failed? Are peo- people share is remarkably similar to theshift to adding 50 cents every few days. A ple falling ill? Before we can act, we need data we collect through household sur-month later, they begin defaulting on mo- more information. veys. They talk about searching for workbile phone repayments of the microloans and not finding it. They talk about gettingthey worked hard to qualify for.17
    • sick, and they list their symptoms. They Lastly, at Global Pulse, we recognize thattalk about what they pay for food and fuel, protection of the vulnerable must includeand about how they are struggling to get protection of their privacy.by with less income than before. So we are collaborating with privacy ex-Across the globe, millions of people are perts around the world to develop guide-sharing information relevant to develop- lines for safely analyzing and using thisment. Real-time data presents us with op- real-time data.portunities we could never have imaginedeven a few years ago. When working at the country level, Global Pulse will use data only in accordanceGLOBAL PULSE SEES IN with national privacy laws, and we have al- ready adopted a internal policy of workingTHESE OPPORTUNITIES only with aggregated data that contains noTHE PROMISE OF A personally identifiable information.FUTURE IN WHICHDEVELOPMENT IS BOTHMORE AGILE AND MORERESILIENT.Yet these opportunities also raise impor-tant questions:What signals do vulnerable communitiesleave behind in real-time data?What tools can help us catch those signalsearly, and how do we separate signal fromnoise?How can we use this data to protect vul-nerable communities from harm, whilealso protecting individual privacy? 18
    • Ms.MAKENA WALKERPartnerships Advisor, Global PulseA t Global Pulse, we recognize that Pulse Labs are physical spaces for open In addition, labs will work with the private the only way to test innovative innovation. sector, to develop ways to safely share and ideas and research is through a analyze data. hands-on approach, using local In these spaces, new tools and approachesexpertise. for using real time data will be researched, Pulse labs bring together government, the develop, tested and scaled. United Nations Country Team, academia, and the private sector.PULSE LABS ARE A Pulse Lab New York will serve as a hub forNETWORK OF COUNTRY- sharing learning and experiences amongst By connecting government to networks ofLEVEL INNOVATION labs and partners. technical resources, labs can complement and enhance existing monitoring systems.CENTERS, ESTABLISHED In the Pulse Lab facilities, new indicators In this way, labs are not establishing a par-IN PARTNERSHIP in real time data will be found, that com- allel structure.WITH MEMBER STATE plement the indicators that governments already track. Once identified, tested andGOVERNMENTS. verified, the most robust indicators dis- covered could be monitored on an ongo- ing basis to contribute to government and United Nations programmes at the coun- try level.19
    • IN 2011, GLOBAL PULSEANNOUNCED THAT THEGOVERNMENTS OFINDONESIA AND UGANDA GOVERNMENT UNITED NATIONSWILL HOST THE FIRSTPULSE LABS, IN ASIA ANDAFRICA. BOTH LABS WILLOPEN IN 2012. PRIVATE SECTOR FOUNDATIONS AND CORPORATIONS AND NGOs OPEN SOURCE LOCAL TECHNOLOGISTS ENTREPRENEURS COMMUNITY ACADEMIA GROUPS 20
    • Long-term benefits: Short-term benefits: Principles of Pulse LabsLABS CONTRIBUTE TO A In the more immediate term, Pulse Labs allow local government institutions and Labs will be different in every country because they will be designed in the wayBETTER UNDERSTANDING United Nations partners to: that best serves a country’s specific needsOF CHANGES IN Identify critical information gaps - and priorities.A POPULATIONS’ and time lag - in understanding the needs of vulnerable populations Labs will derive their value from being col-WELLBEING. THIS COULD Gain access to new sources of real- laborative and multi-disciplinary, as thisHELP THE PUBLIC time information allows government to access diverse andSECTOR TO MORE Gain the tools needed for real-time information analysis new sets of expertise.EFFECTIVELY TARGET Analyze digital data Partnerships with private sector data pro-SCARCE RESOURCES, Discover real time proxy indicators viders will also be established. For exam-AND ACT WHEN THERE Develop local capacity to understand, and use new data-sources and tech- ple, working with mobile service providers would allow the Pulse Labs to be a brokerIS STILL TIME TO nologies of new sources of information that couldPREVENT DEVELOPMENT Demonstrate regional and global lead- point to emerging vulnerabilities within aREVERSALS. ership in innovation. population.21
    • 22
    • INDONESIA PULSE LAB Why is real time data important?23
    • BEATE TRANKMANN UNDP COUNTRY DIRECTOR, INDONESIA “Indonesia already has response systems in place. What they don’t have is the real time data that informs the mitigation strategies to crises. Informing timely responses— that is where the Global Pulse initiative can make a very important contribution.” EL-MOSTAFA BENLAMLIH UN RESIDENT COORDINATOR, INDONESIA “Global Pulse will bring a new kind of ap- proach to making surveys and tracking vulnerability using social media, informa- tion technology and usng mobile phones. And this is very important, because up un- til now, government can track these vulner- abilities, but it has not been on a real time basis.” JACQUI DE LACEY HEAD OF AUSAID, INDONESIA “Indonesia is the perfect place for a Pulse Lab. First, there are 110 million people that live on less than $2 a day. So even minor shocks can have catastrophic impacts here.DR. LUKITA DINARSYAH TUWOVICE MINISTER, STATE MINISTRY OF NATIONAL Second, the Indonesian government is re- ally committed to improving the lives ofDEVELOPMENT PLANNING, INDONESIA the poor and making them more resilient“The nature of today’s crisis are changing. Crisis have become to shock.more complex and often intersect with local emergencies,which compound the impact on the poor and the vulnerable. Third, Indonesia has one of the highest up-The data that we have is very much behind; I believe if we takes of new media anywhere in the world,know where and what household is impacted we can quickly so there is a rich source of data and infor-design a policy response to the affected household. This type mation for the pulse lab to draw on.of high-frequency data becomes very important to protectthe population from the adverse impact of global crisis.” The policy environment here is really recep- tive to new ideas and new information.” 24
    • UGANDA PULSE LAB An emerging technology hub.25
    • Uganda is an emerging technology hub in the region and has built a grow- ing community of technology users and innovators. Kampala has been se- lected as the first Pulse Lab in Africa. Here are the words of team members and leaders in Uganda as they antici- pate the new levels of insight and so- lutions from Pulse Lab Uganda 2012 PETER WAKABI WAISA MANAGER ICT / NATIONAL PLANNING AUTHORITY, UGANDA “There are sources of data that we haven’t been able to reach, but with your mecha- nisms we will be able to reach them. So each problem can be looked at from different perspectives, and that allows the solution to be much better than normally it would be” MICHAEL NIYITEGEKA HEAD OF CORPORATE AFFAIRS,DR. RUHAKANA RUGUNDA MAKERERE UNIVERSITY, FACULTY OF COMPUTING AND INFORMATICS TECH-MINISTRY OF INFORMATION, COMMERCE ANDTECHNOLOGY, UGANDA NOLOGY, UGANDA “We have brilliant students whose ability“The government of Uganda and the civil society would like to conceptualize is largely constrained byto take maximum advantage of Pulse Lab Uganda to get the local environment. Working with thereal-time information in dealing with problems that affect UN Global Pulse facility... you can identifythe ordinary people in our country. We will be able to... Take the small quality students... even membersappropriate measures and even preventive measures so that of staff who have the competency to bringwe can deliver better services for our people.” value to the Pulse lab.” 26
    • Ms.ZAZIE SCHAFERDeputy Director, Global PulseO ver the past year, the Global Pulse And second, the data sources available team has been working with part- were often too aggregated, or too old, to ners from UN Agencies, academia give us detailed insights. and the private sector on some ini-tial research projects. THIS IS WHERE WEFirst we looked back at the global eco- BELIEVE THAT DIGITALnomic crisis, as we wanted to get a better SIGNALS COULD HELP. Summary of Findingsunderstanding of how it actually played These research projects show that itout in peoples’ lives. To explore this hypothesis, Global Pulse is possible to: teamed up with data experts from the pri- Conduct simultaneous rapidFor that, we partnered with research teams vate sector and academic institutions to global mobile surveys at massivefrom 11 UN agencies to look at the crisis conduct 5 research projects. scale;through the lens of 8 different sectors in Gain insight into food price infla-38 countries. tion, day-by-day; Reveal emerging trends in newsThe findings painted a diverse picture and during fast changing crises;two high level themes stood out: Use social media for a deeper understanding of unemployment;First, and not surprisingly, it is really dif- andficult to establish causalities between Leverage Twitter as a real timeglobal shocks and local impacts. indicator of public concerns.27
    • 28
    • RESEARCH PROJECT 1 Taking a Global Snapshot of Wellbeing through Mobile Phones RESEARCH PARTNER: JANASurvey taking is a fundamental tool of We started simple. In August of this year, This exercise illustrated the feasibility ofgovernments and the United Nations, we began a mobile survey around the utilizing mobile phones as a lightweightand our first research project tested new globe asking a few questions related to tool for governments and UN agenciesways to undertake surveys more rapidly people’s well-being. to ask a few targeted questions beforeand with less expense. And not just in a designing more complex and expensivesingle country, but simultaneously acrossthe world. OVER 3 MONTHS, WE household surveys. RECEIVED ALMOST DR NATHAN EAGLE, JANAThe tool for these new global surveys? Mo- 90,000 ANSWERS FROM “My hope is that this is very complemen-bile phones. A powerful platform to collectdata with a reach, speed and economy of OVER 8,000 VOLUNTEER tary to what people are already doing, and already doing well. Being able to get ascale that was unimaginable even a few RESPONDENTS IN 30 handle on the pulse of their citizens is anyears ago. COUNTRIES. important thing and it is something that a lot of governments take very seriously andFor this research project, we collaborated We asked questions like “were you sick do reasonably well. My hope is that withwith Jana, a company with a mobile phone last week?”, “How many days did you work this new set of data, these new insights,data collection network that reaches more last week?” and more subjectively, “how this new window into the behavior of thethan 2.1 billion people in more than 70 did you feel last week?”. aggregate population, it will let them docountries. their job even better.”29
    • 30
    • RESEARCH PROJECT 2 Tracking Commodity Prices by Monitoring Online Prices RESEARCH PARTNER: PRICE STATSNow, let us look at a different type of data:online prices of bread. We wanted to find ONLINE BREAD PRICES This allows us to identify spikes in food prices when they happen. And it helps usout whether online commodity prices cor- DO INDEED FOLLOW track inflationary trends earlier and withrelate with actual price changes on the A SIMILAR TREND AS much greater granularity.street. OFFICIAL PRICES FOR PILAR IGLESIAS, PRICE STATSIf they do, this could help policy makers BREAD. THE DIFFERENCE “It’s interesting because, although in thegain important insights into food price in- IS THAT ONLINE PRICES developing world, it’s true that the numberflation, day-by-day, not month-by-month. CAN BE OBTAINED DAILY of online transaction is not as large, a lot of retailers use online websites as a way toFor this research, Global Pulse partnered WHILST CONSUMER promote their offline prices so it’s a greatwith Price Stats, an economics and tech- PRICE INDICES IN advertising mechanism and just looking atnology company that each day tracks theprices of 5 million products advertised MOST COUNTRIES ARE those retailers, we can have a lot of infor- mation about their offline prices. So youonline. ONLY PUBLISHED ON A don’t need to have a lot of ecommerce MONTHLY BASIS. going on what you need is to have a lotWe looked at the daily price of bread in 6 of retailers that offer retail information inLatin American countries over the last two their websites.”years. And what we found was interesting.31
    • URUGUAY 1.4 1.2 Uruguay Bread Index (daily data) 1.0 Uruguay Bread & Cereals CPI (monthly data) 10 10 10 11 11 11 ch uly ov ch uly ov 1N 1N ar ar 1J 1J A strong correlation was found between1M 1M the Uruguay Bread Index as generated from online prices, and the official Uru- guay Bread & Cereals CPI. The difference is, that the online prices can be tracked daily, while the CPI is a monthly report. 32
    • RESEARCH PROJECT 3 Mining News for Emerging Trends RESEARCH PARTNER: COMPLEX SYSTEMS INSTITUTE OF PARISTENS OF THOUSANDS OF Through tracking emerging news trends online, we analyzed how Francophone This allowed us not only to track these issues over time, but also to show theirONLINE NEWS ARTICLES media reported on food security issues location.ARE GENERATED AROUND over the past 8 years, using the followingTHE WORLD EVERY DAY - method: For example, this graphic represents how the global economic crisis unfolded. YouFAR MORE THAN ANY OF First we identified a set of key words and can see that as it gained momentum, cov-US HAVE TIME TO READ phrases related to food security. erage shifted from a focus on humanitar- ian issues to food price volatility. Then itOur next research project explored wheth- We then retrieved 20,000 related articles changed to social unrest. Interestingly,er we can track important thematic shifts published between 2004 and 2011. The children’s vulnerability has been a topic ofin global attention through mining media contents of these articles were analyzed interest throughout the crisis.articles. and organized in thematic clusters, which can be traced back to specific news ar-For this, we partnered with a consortium ticles or aggregated in a big picture rep-of French research centers and universi- resentation.ties led by the Complex Systems Instituteof Paris Ile-de-France.33
    • This visualization shows the prevalence of different topics in the news. Interestingly, childrens vulnerability, is a topic of inter- est throughout the economic crisis. The Financial CrisisHumanitarian Issues Food Price Volatility Social Unrest Children’s Vulnerability 34
    • RESEARCH PROJECT 4 Understanding Unemployment through the Lens of Social Media RESEARCH PARTNER: SAS INSTITUTEIn our next research project, we investi- We compared the mood expressed in these What this initial research shows is that on-gated what online forums and blogs could conversations with official unemployment line conversations could potentially enrichtell us about unemployment. statistics. In Ireland, we found that fluc- official unemployment statistics. tuations in the anxiety score mirrored fluc-We explored two questions: tuations in the unemployment rate. I-SAH HSIEH, SAS INSTITUTE “Researchers have developed a way toCAN ONLINE Interestingly, the changes in the anxiety score preceded changes in the unemploy- take something qualitative like social me- dia and quantify or score it based uponCONVERSATIONS PROVIDE ment rate. the topics discussed and the feelings ex-AN EARLY INDICATOR OF pressed in those topics. And once some-IMPENDING JOB LOSSES, In the US, the same was true for conversa- tions expressing feelings of anger. thing is quantified, we can use the same universe of statistics tools that are provid-AND CAN THEY HELP ed to national statistics offices to analyzePOLICY MAKERS ENRICH Our analysis also revealed that people millions of data points.”THEIR UNDERSTANDING talk extensively about the impacts of job losses.OF HOW COMMUNITIESCOPE? In Ireland, we noticed an increase in the number of conversations about affordabil-For this research project, Global Pulse ity of housing several months after a risepartnered with the SAS Institute. in the unemployment rate.We started with online conversations in In the US, people talk about transporta-Ireland and the US, where we knew people tion: losing their car, having to downgradewere posting frequently about unemploy- to a smaller car or relying more on publicment. transportation.35
    • ONLINE DISCUSSIONS & UNEMPLOYMENT IRELAND Increase in confusion expressed in online discussions Increase in cancellationsIncrease in anxiety expressed of travel bookings in online discussions ? 5 MONTHS 3 MONTHS 3 MONTHS BEFORE BEFORE LATER Online discussions in Ireland showed an increase in anxiety, 5 months before unemployment; an increase in confusion, 3 months before unemployment; and an increase in cancellation of travel 3 months after unemployment. 36
    • RESEARCH PROJECT 5 Twitter as a Barometer of Food Security Perception in Indonesia RESEARCH PARTNER: CRIMSON HEXAGONThe final project looked at whether Twit- When we looked at tweets specificallyter could provide real-time insights into about rice, we in fact found that theseissues that people are concerned about. conversations closely mirror official foodAnd it turns out, if you want to study Twit-ter, there is no better place to do this thanIndonesia, whose residents tweet nearly 1and a half million times a day.For this research project, Global Pulse 500,000partnered with Crimson Hexagon, a com-pany that analyses social media. 400,000 300,000We wanted to know whether people tweetabout food. And yes! Indonesians clearly 200,000use Twitter to talk about food. 100,000We found more than 14 million tweets re-lated to food over the past 16 months andthe number of tweets can be correlated JULY 2010 NOVEMBER 2010 MARCH 2010 JULY 2011 OCTOBER 2011with real events. The number of tweets per week that areFor instance, you see an increase in peo- related to food in Indonesia.ple talking about food at the beginning ofRamadan, followed by a dip at the end.37
    • inflation statistics. 18,000 TWEETS ABOUT THE PRICE OF RICE 16,000MEGAN COSTELLO, CRIMSON HEXAGON 14,000“So I think what we’re really hoping to 12,000learn is if there’s an additional strategy we of rice Tweets about the price 10,000can employ to get some insight into what 8,000might be causing pressure or stress on 6,000people in their daily lives in a way that’s 4,000more real-time it’s not next quarter whenthe economic indicators come out it this 11 0 11 1 1 1 20 20 20 20afternoon or tomorrow.” AY P N T OC SE JA M OFFICIAL FOOD PRICE INFLATION 3 2 1 0 -1 -2 -3 11 10 11 11 20 20 20 20 AY P N T OC SE JA M When analyzing Indonesian tweets about rice, a strong correlation was found between the number of tweets and the fluctuations of rice price according to of- ficial food inflation statistics. 38
    • TECHNOLOGYHUNCHWORKSA social network designed to support hypothesisformation and evidence-based decision-making.39
    • ROBERT KIRKPATRICK: Long after a full-blown crisis emerges, So at Global Pulse, we began to ask, whatA experts usually look back and admit that if you could find a way to increase the long with research into data, there were already warning signs emerging chances that these experts became aware Global Pulse is also working with early on. Yet the signals were weak, and of each other’s hunches much sooner than technology experts to make sure the evidence was fragmentary. they do today? All over the world, people that government have the tools use social networks to stay up to date withthey need to harness real-time data. In Months in advance, development experts each other.assembling a toolkit for use in our Pulse in different disciplines often have a hunchLabs, however, we found a few gaps. Not — a professional instinct — that some- We wondered, could we use social net-all of the required tools exist. Here, we thing is happening, and they quietly begin works to connect experts working on re-saw an opportunity to innovate. to collect evidence. One notices subtle lated hunches, and create a safe space changes in weather patterns. Another where they could share related evidenceAs you know, we live in a complex world starts tracking migration flows. Another long before they are willing to take it pub-filled with weak signals. Long before a cri- sees certain words and themes appearing lic?sis emerges, there are signals that begin in news stories. Another sees changes into appear in the background noise. Over how people use their phones. If we could connect the people and thetime, these signals grow in strength, until data sooner, it could help us understandthe hard evidence is visible to all, and the But at this point, none of the experts has what is happening while there is still timealarm is sounded publicly. Unfortunately, more than a few pieces of the puzzle, so to prevent harm. It’s not an analyticalby this point, it’s usually too late to act, no one is willing to sound the alarm pub- challenge. It’s a social networking chal-and the harm has already been done. licly. And more often that not, they aren’t lenge. even aware that one another are beginningYet there’s another story here, and it’s one to converge on a common hypothesis.that caught our attention at Global Pulse. 40
    • WE HAVE BEGUNBUILDING A NEW STAGE 1: CREATE A NEW HUNCH +TECHNOLOGY TOOLCALLED HUNCHWORKS -A SOCIAL NETWORK FORFORMING HYPOTHESESAND MAKING EVIDENCE-BASED DECISIONS.HunchWorks allows experts in govern-ment, UN agencies, and academia tomonitor streams of data for digital signals,create hypotheses, share them privatelywith informal networks of colleagues,gather sufficient evidence for verification,and recommend a course of action to de-cision-makers. STAGE 2: SHARE WITHIN TRUSTED NETWORKSIn December of 2010, we brought togethermore than 100 international developmentand technology experts in to understandwhat kind of tool was actually needed, andwhat building blocks were already avail-able.Then we worked with professional technol-ogy designers who volunteered their timeon nights and weekends to help us createa blueprint for HunchWorks. They createdmore than 300 pages of exciting designs.41
    • STAGE 3: VERIFY Today, the basic prototype of HunchWorks is being built. These screenshots will give you an idea of how it works: Stage 1 is to create a new hunch. It is possible to asso- ciate a hunch with one or more locations, and to attach different kinds of evidence, such as data, news articles, or photos. In Stage 2, you can chose who to share a hunch with. Most of the experts we have consulted with told us that in the early days, they want to keep hunches private, visible only to a few trusted colleagues. The HunchWorks system allows for that. In Stage 3, you can see comments that colleagues have shared about your hunch, and see an overview of the evidence asso- ciated with a hunch. Different experts canSTAGE 4: TAKE ACTION share their levels of confidence in individ- ual pieces, giving any individual hunch an overall confidence score. Finally, in Stage 4, you are presented with the option to package up the evidence and generate a report. If you are interested in learning more about our progress on HunchWorks, information is available on the Global Pulse website. 42
    • ROADMAP & CONCLUSIONSAfter reviewing how we are approaching with cutting-edge data mining experts. ing Framework, a set of guidelines on newGlobal Pulse’s mission, and what the proj- indicators, analytical techniques, tool andect has accomplished over the past year, Pulse Lab New York. Our New York Pulse processes that any government may beginlet us now turn to a review of our top pri- Lab will continue to function as the hub adopting into standard practice.orities for 2012 and 2013. of the Pulse Lab Network, providing coun- try-level Pulse Labs with foundational TechnologyPulse Labs tools and methods they may build upon Hunchworks. In JanuaryPulse Lab Network. In and developing strategic partnerships with of 2012, Global Pulse will2012, Pulse Lab Kampala private sector and academia. make the first version ofand Pulse Lab Uganda will HunchWorks available foropen their doors. With your The Long View. Pulse Labs are a bridge on testing by a limited num-support, we will open two our journey to a more agile form of inter- ber of experts in govern- TECHNOLOGYmore Pulse Labs in 2013. national development. In the years ahead, ment, UN agencies, and PULSE LABS Global Pulse will measure success by the academia. A public version will be avail-The Pulse Lab Network will volume of innovation it fosters and by the able for testing in August. HunchWorksbe ramping up research efforts significant- velocity at which those innovations are ad- will be used by the Labs as a core tool inly over the next two years. opted into practice. their toolkit for analyzing real-time data.Impact Monitoring 2.0. To facilitate this Researchresearch, with the support of the MDG Real-Time MonitoringFund, Global Pulse will be establishing a Framework. By the end ofnew fund in 2012, called Impact Monitor- 2013, Global Pulse willing 2.0. This fund is an innovative mech- publish the first versionanism for joint research that will match of its Real-Time Monitor-interested governments and UN agencies RESEARCH43
    • Data Access organizations to begin establishing theseFinally, at the beginning of today’s pre- public/private data-sharing partnerships.sentation, you heard Mr. Hirshberg and The potential here is enormous. Can youDr. Weigend describe real-time data as a imagine a world in which public and pri-core global resource. Yet one challenge vate sector join forces to create a publicwe face is that much of this data is not commons of real-time data, to accelerateaccessible. It is collected by companies international development and strengthenthat use it to competitive advantage, and our collective resilience to shocks?it isn’t shared publicly. You have now heard about how GlobalSo Global Pulse has begun engaging with Pulse is working to achieve its vision. Itprivate sector companies about an idea has been our great pleasure to demon-we call “data philanthropy.” Data philan- strate to you how compelling the potentialthropy means that private sector compa- is for new data to transform our work. Wenies share analysis of their data to support look forward to engaging with you, espe-international development. The compa- cially in the establishment of our networknies we have spoken with are willing and of Pulse Labs, in the months and yearsinterested. ahead.In 2012, Global Pulse will begin assem-bling a consortium of private sector com-panies, universities, and and international 44
    • Global Pulse TeamRobert Kirkpatrick, Director Research Partners & CollaboratorsZazie Schafer, Deputy Director SAS InstituteMakena Walker, Partnerships Advisor JanaCassandra Hendricks, Project Associate Crimson HexagonAnoush Tatevossian, Strategic Communications Officer PriceStatsMiguel Luengo-Oroz, Lead Data Scientist Complex Systems Institute of Paris Ile-de- FranceEva Kaplan, Programme Specialist Institut Francilien Recherche Innovation Société (IFRIS)Emmanuel Letouzé, Development EconomistSara Farmer, Chief Platform Architect Data VisualizationsChris Van Der Walt, Strategic Communications Advisor Nik Hanselmann for EpidemicIQAdam McKaig, Lead Developer MIT Senseable City LabPatrick Adengo, Innovation Officer, Kampala Complex Systems Institute of ParisChristopher Fabian, Innovation Advisor (UNICEF) Map KiberaUN Agency Partners Book EditorsUNDP, UNICEF, WFP, UNOOSA, UNEP, UNECA, UNESCO, Peter HirshbergUNODC, UNWTO, ILO, UNFPA Christine OutramMember State Partners PhotographyIndonesia, Uganda, UK, Sweden, Australia Photos ©2011 Rick Smolan/Against All Odds ProductionsStrategy & Innovation Advisors Book DesignPeter Hirshberg, Annenberg Center on Communication, Leader- Nicole Aptekarship & Policy Christine OutramChristine Outram, City Innovation Group Graphic Design Chris Van Der Walt45
    • www.unglobalpulse.org@unglobalpulsefacebook.com/unglobalpulse 46