Presentation by Aklilu Dinkneh Teklesadik (Red Cross 510 Team) at the Data Science Symposium 2018, during Delft Software Days - Edition 2018. Thursday 15 November 2018, Delft.
4. Shape the future of humanitarian aid by converting data into
understanding, and put it in the hands of humanitarian relief workers,
decision makers and people affected, so that they can better prepare
for and cope with disasters and crises.
MISSION
5. Shape the future of humanitarian aid by converting data into
understanding, and put it in the hands of humanitarian relief workers,
decision makers and people affected, so that they can better prepare
for and cope with disasters and crises.
MISSION
6. Shape the future of humanitarian aid by converting data into
understanding, and put it in the hands of humanitarian relief workers,
decision makers and people affected, so that they can better prepare
for and cope with disasters and crises.
MISSION
25. DATA COLLECTION: VULNERABILITY
wall material per municipality
.
WALL TYPE
ROOF TYPE wood Concrete Mud bricks
Thatch CLASS 1 CLASS 5 CLASS 2
Corrugated iron CLASS 3 CLASS 6 CLASS 4
26. Travelling time to the nearest hospital.(coping
capacity indicator)
Which is calculated based on OSM data.
DATA COLLECTION: COPING CAPACITY
27. Damage and Needs Assessment reports
Digital newspaper repositories
DREFs
DATA COLLECTION: IMPACT DATA
28. • Data from local and national level
• Damage and Needs Assessments, such as from
National Disaster Management agencies or NGOs
• Digital newspaper repositories
• DREFs
• Data from global repositories (often derived from
national databases)
• EM-DAT
• Desinventar
• Preventionweb
• Humanitarian Data Exchange
DATA COLLECTION: IMPACT DATA
35. Humanitarian finance is available mainly when a disaster strikes and suffering is
almost guaranteed.
But climate-related risks are rising
Challenge
Climate
change
Population Poverty
Exposure
Hazard
(weather and
climate
events)
Disaster
Risk
Vulnerability &
Coping Capacity
FORECAST BASED FINACNING - IBF
• Introduction
• Data collecten
• Data Integration
• IBF
• Other DS activities
36. Forecast-based financing (FBF) releases humanitarian funding based on
forecast information TO TAKE PREDEFINED ACTIONS TO reduce risks,
The Innovation
We can forecast climate-related risks (with uncertainty) with a lead time
Humanitarian actions could be implemented in this lead time window
Opportunity
FORECAST BASED FINACNING - IBF
• Introduction
• Data collecten
• Data Integration
• IBF
• Other DS activities
38. IMPACT BASED FORECASTING- IBF
Input (explanatory variables) Output (loss and damage)
Composite index approach (overlay) Based on experience, usually in
relation to one input indicator (wind
speed, water level) at specific
locations.
Often estimates, no absolute values
Elementary modelling (rule-based) Data analysis is done to e.g.
determine thresholds. Simple
damage-hazard curves with one
input variable.
Often for infrastructural damage.
Statistical modelling (without relying
on rule-based)
Multiple indicators also for e.g.
urban vs rural.
Also for e.g. crop damage modelling
41. Damage-hazard curve: identify underlying
causes of impacts (vulnerabilities)
Destruction of Houses
Bad quality construction materials (wall and roof)
Poverty
Vulnerability Indicators
IMPACT BASED FORECASTING- IBF
ELEMENTARY MODELLING
42. • Collect for historical typhoons data on
loss and damage (output) & several
explanatory variables such as wind
speed, wall/rooftypes (input)
• Make statistical model to predict
damage and improve model
performance
• For upcoming typhoon collect same
inputs (forecasted wind speed) and and
apply model to predict output (damage)
Note the change of
forecasted typhoon track in
12hrs time. Damage
prediction can only be as
good as the weather forecast!
IMPACT BASED FORECASTING- IBF
STATISTICAL MODELLING/MACHINE LEARNING