Presentation of an overview of my PhD research at "Beyond research - Patways to impact" RMIT Conference, Melbourne, 21,22,23 February 2017

1. Deriving near real time geospatial footprints
of crisis events
combining information from social media,
crowdsourced and authoritative data
Valentina Cerutti
valentina.cerutti@rmit.edu.au
School of Science

2. Introduction
Valentina Cerutti - ‘Beyond Research – Pathways to Impact’ 2017 RMIT Conference – 22/02/2017 2
000
DISASTER
EMERGENCY SERVICES
?
? ?
?
SOCIAL SENSORS
AUTHORITATIVE DATA
 Geospatial data
 Sensors Networks
 Satellite Images
 Models
 …
 VGI
 Crowdsourced data
 Social Media
+ Real time
+ Cost effective
- Trust
- Credibility
- Quality
- Data shadow
- Unstructured
- Noisy
- …

3. Geospatial footprint
Valentina Cerutti - ‘Beyond Research – Pathways to Impact’ 2017 RMIT Conference – 22/02/2017 3
Analysis of semantic, spatial and
temporal characteristics of social
media data
Combination with ancillary
authoritative data  geographic
contextualization
Aggregation and representation of
disaster-related information as a
geospatial footprint
Near real time
Automatic procedure
!
Why geospatial footprint?
• Approximate imprecise areal regions from point locations
• Gain insight on the (spatio-temporal) dynamics of events
• Help synthesize and contextualize various type of geographic information
• Visualize current and future (potential) impact of the event

4. Exploratory visual analysis of Twitter data
Comparison with ground-truth data
Valentina Cerutti - ‘Beyond Research – Pathways to Impact’ 2017 RMIT Conference – 22/02/2017 4
18-19 Nov. 2013
V. Cerutti, G.Fuchs, G. Andrienko, N. Andrienko, F. Ostermann. Identification of disaster-affected areas using exploratory visual analysis of georeferenced Tweets: application to a flood event. 19th AGILE
Conference on Geographic Information Science 2016, Helsinki, Finland
Georeferenced
Tweets
Data mining and visual analysis
Real-time
simulations
OPTICS Distance-bounded
event clustering
Data-driven territory
tessellation False positive
Application case:
flood in Sardinia (Italy)
18-19 Nov. 2013
Flood-related Tweets

5. Reproducible research
Valentina Cerutti - ‘Beyond Research – Pathways to Impact’ 2017 RMIT Conference – 22/02/2017 5
use scientific workflows to support spatial data analytics
for building geospatial footprints of events from crowdsourced data
visual language and
self documenting analytics
conceptualize and manage
the analysis process
Research reproducibilitycreation and reuse of analysis tasks
Qualitative + quantitative evaluation

6. GEOSAFE project
Valentina Cerutti - ‘Beyond Research – Pathways to Impact’ 2017 RMIT Conference – 22/02/2017 6
 Research application case: Bushfire in Australia
 EU-funded project with several partners (Universities, stakeholders and end-users) in Europe
and Australia
 2016-2020
Geospatial based Environment for Optimisation Systems Addressing Fire Emergencies
 Development of innovative tools for fire management
 Develop and assess global or semi-global dynamic tools for:
fire suppression, lives/goods protection and implementation
and training

7. Future research and challenges
 Combination of heterogenous data sources
 Integration of scientific workflows to create geospatial footprints from heterogeneous data
into disaster response mechanisms
 Representation of current and potential impacts
 Near real time application
Valentina Cerutti - ‘Beyond Research – Pathways to Impact’ 2017 RMIT Conference – 22/02/2017 7

8. Deriving near real time geospatial footprints
of crisis events
combining information from social media,
crowdsourced and authoritative data
Valentina Cerutti
valentina.cerutti@rmit.edu.au
Thanks for your attention!