Transcript
Page 1: How to Predict a Serial Arsonist

How to Predict a Serial ArsonistEva Ljungkvist Björn Granvik

Page 2: How to Predict a Serial Arsonist

Dan Granvik !former County Sheriff’s Department Special Unit focusing on Organized Crime & Repeat Offenders

Eva Ljungkvist !Fire protection engineer Arson investigator i.e. Fire-nerd

Björn Granvik !Civil Engineer Programmer i.e. IT-nerd

Page 3: How to Predict a Serial Arsonist

• Like most peoplehabits, patterns, …

• Modus operandi

• Evidence?

The Serial Arsonist

Page 4: How to Predict a Serial Arsonist

Copper

Page 5: How to Predict a Serial Arsonist

High risk?

Page 6: How to Predict a Serial Arsonist
Page 7: How to Predict a Serial Arsonist

© Guillaume Baviere, Flickr

Page 8: How to Predict a Serial Arsonist
Page 9: How to Predict a Serial Arsonist

Method Triangulation

Page 10: How to Predict a Serial Arsonist

Powered by Fika

Page 11: How to Predict a Serial Arsonist

Timings

Device Primed Device Ignited Fire Discovered

Case 33 23:19

Case 34 07:12

Page 12: How to Predict a Serial Arsonist

In the Head of Dan

© TheKarenD, Flickr

Page 13: How to Predict a Serial Arsonist

In the head of an Arsonist?

• Data is too complex

• …and hidden

• Power in relationships

© Jenny Spadafora, Flickr

Page 14: How to Predict a Serial Arsonist

A Graph• Node

”noun”

• Relationship”structure”

• Properties

Aname: ”Björn” headsize: 60

technology: ”Neo4j”

strength: 10

B

LIKES

C USES

name: ”Eva”

USES

Dproject: ”Helmuth”

USED_IN

HAS_MEMBER

Page 15: How to Predict a Serial Arsonist

Pattern Matching

A

B C

Page 16: How to Predict a Serial Arsonist

Pattern Matching

Page 17: How to Predict a Serial Arsonist

An example

Hospital

Arson Device

name: ”Heimlich Hospital” adress: ”21 Prune Street" city: ”Snicket Town"

lives_at House A

name: ”Hus A” adress: ”22 Prune Street" city: ”Snicket Town"

description: ”Toarulle”

Fire

description: ”Big poff” with

happened_at

Timings

name: ”Count Olaf”

next_to

started

Page 18: How to Predict a Serial Arsonist

Example Queries◦ Do we have a pattern, a serial arsonist?!◦ What new targets? High vs. low risk?!◦ Where and when should we apply preventive

measures?

Page 19: How to Predict a Serial Arsonist

How to Predict?

◦ Crossover ◦ Support complex

world descriptions ◦ Usability

© Moyan Brenn, Flickr

Page 20: How to Predict a Serial Arsonist

Results so far◦ Weather conditions!◦ Time frame!◦ Trends of arson !◦ Location & arson device!◦ High risk objects & the need of fire protection!◦ Worst case scenario & the need of resources!◦ Physical security & immediate surroundings

Serial

Page 21: How to Predict a Serial Arsonist

Eva Ljungkvist!

!

[email protected]

http://evasbrandblogg.se/

Björn Granvik!

!

[email protected]

@bjorngranvik

http://bit.ly/bjorngranvikblog

Thanks!


Recommended