Big data and development

Big data & development
The whats, whos & whys
SIMONE SALA - @HEREISSIMONE - WWW.SIMONESALA.IT
* ASSOCIATE DIRECTOR OF THE DR. STEVE CHAN CENTER FOR SENSEMAKING,
AIRS (HAWAI’I PACIFIC UNIVERSITY AND SWANSEA UNIVERSITY)
* RESEARCH AFFILIATE, DATAPOP ALLIANCE
* FELLOW, UNIVERSITY OF MILAN GEOLAB
ICTP – Trieste – March 26, 2015

Big data & development
• What?
• Who & Where?
• Why?

What is big data (for development)?
• Many definitions: is there a right one?
• What is big data (as of today)?
• What is big data in/for development?

What is big data?
Definitions
• Is there a single definition of big data?

What is big data?
Definitions
• No single agreed definition of big data.
• Though a lot of people is talking about big data… isn’t it?

What is big data?
Definitions
Let’s see what data is saying about big data!
Source: http://www.google.com/trends/explore#q=big%20data

What is big data?
Definitions
• No single agreed definition of big data.
• Though a lot of people is talking about big data!
• So is that just hype?

What is big data?
Definitions & Hype
http://www.gartner.com/technology/research/methodologies/hype-cycle.jsp

What is big data?
Definitions & Hype

What is big data?
Back to definitions
• Is there a single definition of big data? There is no single agreed definition of big data.
• Let’s take a look at a few of them…
• “Big data is a broad term for data sets so large or complex that traditional data processing
applications are inadequate. Challenges include analysis, capture, curation, search, sharing,
storage, transfer, visualization, and information privacy. The term often refers simply to the use
of predictive analytics or other certain advanced methods to extract value from data, and
seldom to a particular size of data set.” – Wikipedia

What is big data?
Definitions
• Is there a single definition of big data? There is no single agreed definition of big data.
• Let’s take a look at a few of them…
• “Big data usually includes data sets with sizes beyond the ability of commonly used software tools
to capture, curate, manage, and process data within a tolerable elapsed time.”
- Snijders, C.; Matzat, U.; Reips, U.-D. (2012)
• “Big data is a set of techniques and technologies that require new forms of integration to uncover
large hidden values from large datasets that are diverse, complex, and of a massive scale.”
- Ibrahim Abaker Targio Hashem et al. (2015)

What is big data?
Definitions
• Size matters?
• When is data big enough to talk about big data?

What is big data?
Definitions
• Size matters? When is data
big enough to talk about big data?
• A.k.a. is this big data?

What is big data?
Definitions
• Size matters? When is data big enough to talk about big data?
• The big thing about big data is that we have new data coming from a variety of
machines/people through electronic sources into databases for no statistical inference purposes
• It’s not (only) about size

What is big data in/for development?
• Is this a brand new story?

In short – nope.
CyberSyn,Chile-1971

What is big data and development?
In short – nope.
• So what is different?

• Many reasons why big data is different and matters for development and in developing regions
• Thanks to social media & mobiles we have new (machine-readable) data, generated about and
by people that was compeletely unavailable 10 years ago

New (machine-readable) data
from the developing world

What is big data in/for
development?
New (machine-
readable) data from
the developing world
“Egypt, Russia, the Philippines and 14 other
DCs outpace the U.S. in the proportion of
Internet users who log on to social sites.”
[PEW Research Center]

• Many reasons why big data is different and matters for development and in developing regions
• Thanks to social media & mobiles we have new (machine-readable) data, generated about and
by people that was compeletely unavailable 10 years ago
• It’s the analytics beyond data: we maximized our power to collect + manage/analyze + transmit
data in the past decade, and we will further do it
• The medium community is the message: open data & big data communities behind
deployments
• E.g. Call Data Record vs World Bank database of indicators (again – size is not the key attribute!)

• It is not necessarily a technology-push that enables data-driven innovations in/for
development…

The Toilet Gap
• Here is a story to think about…
• Toilet Gap
• Height of a child below average is among main proxy to describe her/his bad health status
• International policies against malnutrition privileged food security interventions
• Recently, something was found out…

• The Toilet Gap!
• Sanitation is heavily correlated
with children malnutrition
• [Through big data?] new policies
could have been formulated
Source: World Bank
http://goo.gl/qkzj1X

Main data types
3 main types of data being already used
1. Structured datasets – e.g. Call Detail Records (CDRs) collected by mobile phone operators,
metadata capturing mobile phone subscribers’ use of their cell-phones (identification code,
location of the phone tower that routed the call for both caller and receiver, information about
the call)
2. Media content: both in traditional media (e.g. videos, documents, blog posts) and social media.
3. Data from digital sensors – e.g. vegetation indexes from satellite sensors, smart meters to record
water consumption

Who is actually doing «big data for
development»?
• Main actors
• Main experiences

Who is doing that?
Various visions
• Private sector (not just tech corporations…) as a market strategy – e.g. IBM, Monsanto, etc.
• Private sector as a CSR strategy – e.g. Orange Telecom, etc.
• Governments – e.g. Governments of low- and middle-income countries, USAID, etc.
• Inter-governmental organizations – e.g. United Nations, World Bank, etc.
• Academia & research institutes – e.g. DataPop Alliance, Columbia University IDSE, MIT Media
Lab, Université Catholique de Louvain, etc.
• Non profits – e.g. Rockefeller Foundation, Knight Foundation, Bill & Melinda Gates Foundation,
DataKind
• …

Who is doing that?
Clashing (sometimes) narratives
• Big data as panacea VS Big data as evil
• Kenneth Cukier, data editor at The Economist
http://www.economist.com/multimedia?bclid=0&bctid=2380227548001
• Andreas Weigend, former chief scientist at Amazon, introduced the ’data is the new oil’ concept
• E. Morozov “The rise of data and the death of politics”
http://www.theguardian.com/technology/2014/jul/20/rise-of-data-death-of-politics-evgeny-
morozov-algorithmic-regulation
• Danah Boyd and Kate Crawford, “Six provocations for big data”
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1926431

Who is doing that?
• United Nations: Global Pulse (UN’s Big Data innovation lab)
• Launched in 2009
• Aims at developing the capacity to responsibly use new data sources and analytical tools
• Bridge among data scientists, data providers, governments, and development sector
practitioners
• Both evangelization and practice
• Three labs: NYC, Kampala & Jakarta
Source: http://www.unglobalpulse.org

Who is doing that?
• Governments:
• Kenya launched its new Open Data Portal in 2011 - full digital edition of 2009 census, 12 years
of government expenditure data, government household income surveys, location of schools
and health facilities
https://opendata.go.ke/
• The Philippines decided to establish a Big Data Center in 2014 to improve disaster response
https://www.senate.gov.ph/press_release/2014/0625_aquino1.asp
• Mexico leveraged CDRs to analyze the country’s response to the 2009 swine flu outbreak as well
as to strengthen future response to flooding
http://goo.gl/j3w9mA
• Ghana started mining social web to keep track of citizens’ satisfaction on their government

Who is doing that?
United Nations: Global Pulse’s main taxonomy
• Data Exhaust - passively collected transactional data from people’s use of digital services like mobile phones,
purchases, web searches, etc., these digital services create networked sensors of human behavior;
• Open Web Data - Web content such as news media and social media interactions (e.g. blogs, Twitter), news
articles obituaries, ecommerce, job postings; sensor of human intent, sentiments, perceptions, and want.
• Citizen reporting - information actively produced or submitted by citizens through mobile phone-based surveys,
hotlines, user-generated maps, etc
• Physical Sensors - satellite or infrared imagery of changing landscapes, traffic patterns, light emissions, urban
development and topographic changes, etc; remote sensing of changes in human activity

Who is doing that?
United Nations: Global Pulse’s 3 main domains of application
• Early warning - “early detection of anomalies in how populations use digital devices and
services can enable faster response in times of crisis”;
• Digital awareness - “Big Data can paint a fine-grained and current representation of reality
which can inform the design and targeting of programs and policies;”
• Real-time feedback - “monitor a population in real time makes it possible to understand where
policies and programs are failing and make adjustments.”

Who is doing that?
United Nations: Global Pulse’s mining Indonesian tweets to understand food price crises (2013)
• Tweets about the price of rice
vs. actual price of rice in Indonesia

Who is doing that?
• Tweets about the price of beef
vs. actual price of beef in Indonesia

Who is doing that?
• Tweets about the price of chicken
vs. actual price of chicken in Indonesia

Who is doing that?
• Tweets about the price of onions
vs. actual price of onions in Indonesia

Who is doing that?
United Nations: Global Pulse’s analysis of global engagement on climate change via twitter (2014)
• Towards 2014 UN Climate Summit: real-time social media monitor to measure and explore
online discourse about climate change
• Goal: Assess volume & content of tweets about climate change (in 3 languages) across a range of
topics (e.g. economy, energy)

Who is doing that?
Climate March
Climate Summit
US – China
deal on carbon cuts

Who is doing that?
United Nations: Global Pulse’s analysis of global
engagement on climate change via twitter (2014)

Who is doing that?
Climate March
COP 20 in Lima
G7 new global deal
on climate change
?

Who is doing that?
• Towards 2014 UN Climate Summit: real-time social media monitor to measure and explore online
discourse about climate change
• Goal: Assess volume & content of tweets about climate change (in 3 languages) across a range of
topics (e.g. economy, energy)
• Outputs: development of a baseline of engagement by measuring public tweets over time
(highlighiting spikes during the Summit); identification of main sources of information in the social
web; …
• Outcomes: improved ability to compare interest level between topics/regions; improved capacity
to monitor impact of social media impact of climate-related communications (thus higher capacity
to measure awareness, support climate policy decision-making, foster public engagement, …).
• Source: http://www.unglobalpulse.net/climate/

Who is doing that?
United Nations: Global Pulse’s development of food security estimates via mobile phone data &
airtime credit purchases in East Africa (2015)
• Research conducted with UN WFP, Université Catholique de Louvain and Real Impact Analytics
• Comparison of data extracted from airtime credit purchases (or “top-ups”) and mobile phone
activity with nationwide household survey conducted by WFP at the same time.
• Outcome: high correlations between airtime credit purchases and survey results referring to
consumption of several food items – e.g. vitamin-rich vegetables, meat or cereals.
• Source: http://www.unglobalpulse.org/mobile-CDRs-food-security

Who is doing that?
United Nations: Global Pulse’s development of food security estimates via mobile phone data &
airtime credit purchases in East Africa (2015)
Correlation between consumption of foods
and the sum of airtime credit purchases Correlation between mobile phone variables and the
developed Multidimensional Poverty Index

Who is doing that?
United Nations: Global Pulse’s additional activities
• Assessment of indicators of various socioeconomic variables – e.g. unemployment through big
data
Source: http://www.unglobalpulse.org/projects/cansocialmediaminingadddepthunemployment-
statistics
• Feasibility study with USAID to identify barriers to financial inclusion in Kenya via new digital
data sources
Source: http://unglobalpulse.org/Kenyanaccessfinance
• Mining the social web (blog posts, online comments and tweets) with UNICEF to gain insights
into the attitude of parents towards immunization of their children. Leverage the findings to
improve communication with parents to allay their fears.
Source: http://www.unicef.org/ceecis/media_24017.html

Who is doing that?
• United Nations (post2015 agenda)
“Better data and statistics will help governments track progress and make sure their decisions are
evidence-based; they can also strengthen accountability.
This is not just about governments. International agencies, CSOs and the private sector should be
involved.
A true data revolution would draw on existing and new sources of data to fully integrate statistics into
decision making, promote open access to, and use of, data and ensure increased support for statistical
systems.”
U.N. High Level Panel report on the post2015 agenda
Source: http://www.post2015hlp.org/featured/highlevelpanelreleasesrecommendationsforworlds-
nextdevelopmentagenda/
+ World Bank’s work on the same topic: Shanta Devarajan, Africa’s statistical tragedy
http://blogs.worldbank.org/africacan/africa-s-statistical-tragedy

Who is doing that?
Academia & research institutes - Engineering Social Systems lab, Harvard
• Smart (?) slums: Kibera (Nairobi, Kenya)
• Coupled analysis of (a few terabyte of) data gathered from mobiles and information extracted from
national census
• Mobile phone call logs from June 2008 to June 2009
• Development of a growth model of the slum, inferring places of work and migration out of Kibera
• Ultimate goal: make municipalities aware about the places where water pumps & health services will
mostly be needed
Source: http://www.hsph.harvard.edu//ess/

Who is doing that?
Academia & research institutes –
New England Complex Systems Institute
• Model and predict food riots based on
historical data about food prices
Source:
http://necsi.edu/research/social/food_crises.pdf

Who is doing that?
Academia –
Computer Science @ JHU
• Model and predict flu outbreaks in the US
based on tweets
Source: Paul and Dredze (2011)
http://www.cs.jhu.edu/%7Empaul/files/2011.icwsm.twitter_health.pdf

Who is doing that? And where?
Academia – joint research groups for D4D Challenge (2013)
• Exploration and analysis of 2.5 billion calls and SMS
exchange between around 5 million users located in Ivory
Coast over a period of 5 months.
• Analysis of linkages between localized increase and
decrease of calls with major local events – e.g. increased
phone calls correlated with conflict events
Source: Van den Eltzen et al. (2013) http://goo.gl/DNPmg4

• Analysis of linkages between localized increase and
decrease of calls with major local events – e.g. shut-down
of network after the elections in selected areas of the
country

• Analysis of linkages between localized increase and decrease
of calls with major local events – e.g. increased phone calls
correlated with weather anomalies
• Possible way forward:
Optimization of irrigation, logistics, harvesting
Early warning (e.g. calls correlation with weather events,
pests, diseases)
Yield performance (e.g. calls at marketing time in cocoa-
producing region)

Why are we doing this?
Access issues
• New digital divides
• Can we trust data? At what extent can we trust data?

Access issues
• Who are we leaving behind?

Access issues
• Quite a lot of persons, actually.

• Quite a lot of persons, actually.

Access issues
• Even when it is just for a while – it matters
• Sub-Saharan African countries suffer 5-hour power cuts on average every 4 days. This is 25%
more frequent and almost 2x as the overall average of the 135 predominately developing
nations investigated [World Bank's Enterprise Survey]

Access issues
• Access to data is not always granted: researchers’ and policy makers’ ability to access new
sources of data is critical (as much as their ability to properly maintain their old sources)
• Most of new data sources are proprietary

Capacity issues
• Can we actually use all our data?
“The explosion of big data has far-outpaced our ability to make sense of it in poorer nations that
already lack human and technical capacity.” - Claire Melamed, Overseas Development Institute
“The real problem is that governments are swimming in more data than they have ever had but
they lack the capacity in their staff to do anything with it” - Jon Gosier, D8A Group CEO
• Strengthen capacities is critical

Interpretation issues
• Can we trust data ?
“Nobody believes in simulation models except their developers…
Everybody believes in experimental data except who collected them”
- Prof. Gaylon S. Campbell

• At what extent can we trust data ?
• Interpretation issues
• Correlation is not causation
“When you have a billion observations, everything’s significant” – Prof. Hal Varian, Chief
Economist at Google
• Data cannot always speak for itself: expert knowledge is often needed (meaning of data is not
stable, and it is hard to detect bias in data)
“[…] still you will need theory to understand the mechanisms or even to suggest what you might
hope to find in the first place.” Prof. Sascha Becker, University of Warwick

• At what extent can we trust data ?
• Still a long, long way to go…
• Behind every data/algorithm there is a human
• Quality matters: 80% of a data scientist's time spent
cleaning up data
• Good data >>> Big data:
E.g. GFT has over-estimated the prevalence
of flu for 100 out of the last 108 weeks;
it’s been wrong since August 2011.

Ethical issues

Ethical issues
• When Big Data ends and Big Brother begins?
• Are we sure hyper-transparency is always a great thing (e.g. for individuals who belong to a
minority)?
• Risks behind unilateral models of intervention in development settings
• What about the open data dividend concept?
http://www.ictworks.org/2015/03/18/how-can-we-all-profit-from-development-data/
• …
Further information on many of these issues: https://linnettaylor.wordpress.com/

Thank you very much!
Feedback welcome: salas@mit.edu

Big data and development

More Related Content

What's hot (20)

Similar to Big data and development (20)

More from Simone Sala (16)

Recently uploaded (20)

Big data and development

Editor's Notes