peace deals revisited presentation
DESCRIPTION
How can we use big data for peace and security...? Peace Deals Revisited is a project of the Peace Informatics Lab at Leiden University (Campus The Hague)TRANSCRIPT
PEACEDEALSREVISITED
The Peace Informatics Lab is a project of the Centre for Innovation, Campus The Hague (Leiden University)In cooperation with: Leiden Institute for Advanced Computer Science (LIACS) Pax Ludens Foundation Grotius Centre The Hague Centre for Terrorism and Counter-Terrorism (CTC)
VisionStatus Quo
Peace studies today are driven by qualitative or quantitative research. Central to research designs: testing hypotheses about what works in a certain contextSpecific fields of expertise provide the main perspectives (conflict studies, economics, etc.)With the emergence of big data: we can rethink some of these routines.
We want to use big data to generate new hypotheses without a clear-cut disciplinary starting point
WHY BIG DATA
STARTING POINTS
Data analytics is not yet applied for peace studies
Data availability will increase in the coming decade
Proof of concept required for identifying way forward
WHAT IS BIG DATA
APPROACH
Data mining(Yes!)
Statistics(No)
8%
subgroupDISCOVERY
We ‘interview a database’:
algorithms help in finding patterns that are otherwise hard to find.
However:
They do not find causalities.
The patterns have to be interpreted by domain experts.
(from theory and policy practice)
METHODOLOGY
+2y
Looking for correlations
the country context before a peace
agreement was signed
the change in country context after a peace agreement was signed
-2y
ECONOMIC
SOCIAL
POLITICAL
SOCIO- CULTURAL
TECHNOLOGY
LEGAL
ACADEMIC
DEMOGRAPHY
ENVIRONMENTAL
METHODOLOGY
MEDIA
REGULATORY
MILITARY
ECONOMIC
SOCIAL
POLITICAL
SOCIO-CULTURAL
TECHNOLOGY
LEGAL
ACADEMIC
DEMOGRAPHY
ENVIRONMENTAL
REGULATORYMILITARY
+2y-2y
DATA & SOURCES
Current DatabaseData Sources
Open, public access
Global coverage
Long-term coverage
Diverse in focus
1000+ UN SC Resolutions
50+ affected countries
900 variables
8484%% 4343%% 22%%
DATAavailability today (Africa)
peace agreements overTIME
HEAT MAPThe database fusion platform offers insights into the overall correlations.
It points to areas of interest (dark-shaded)
SUBGROUP DISCOVERY
Using CORTANA* the Peace Informatics Lab team build the proof of concept for using big data for peace studies.
* CORTANA is a data mining software from the Leiden Institute for Advanced Computer Science (LIACS)
“We can show if a certain group of peace deals shows a societal change that is significantly
different from the average correlation”
[the following slides are illustrations based on the data thus far collected – to be expanded]
someSUBGROUPS
If textiles and clothing (% of value added manufacturing) is higher than 1.39% at the time of the resolution, then voice and accountability shows a
2.2% increase after 2 years (average is 1.36% decrease)This applies to 143 resolutions, 11 countries
-2 +2y
someSUBGROUPS
If governance effectiveness is lower than -1.6 at the time of the resolution, then the number of fixed telephone lines shows a 17.4%
increase per year after 2 years (average is 2.9% increase)This applies to 119 resolutions, 8 countries
-2 +2y
someSUBGROUPS
If total health expenditure (% of GDP) is lower than 4.7% at the time of the resolution, then voice and accountability shows a 0.2% increase after
2 years (average is 1.36% decrease)This applies to 735 resolutions, 44 countries
-2 +2y
someSUBGROUPS
If average temperature in June is above 31.7 °C at the time of the resolution, then expenditure-side real GDP at chained PPP shows a
10.7% increase after 2 years (average is 3.9% increase)This applies to 123 resolutions, 3 countries
Iraq, Kuwait, Chad
-2 +2y
USEFUL?
These examples give a first impression of the type of correlations that could be uncovered. Meaningful analysis will have to be based on
much larger amounts of data, such as cell phone, twitter, etc.
In 2014/2015 the Peace Informatics Lab will work with a growing group of international partners to bring a sufficiently large dataset online and to
make it available for public use.
The way forwardTOWARDS BIG DATA
Expand the national data
Expand beyond national data
add local peace deals (provincial coverage)
add local data on variables (based on data from NGOs, etc)
add monthly data
create your own data (crowdsourcing)
...
add peace deals
add academic & media coverage
add UN peacekeeping presence
add twitter, cell phone data
add climate data
...