Quantifying human behaviour using online data. Suzy Moat Data Science Lab Behavioural Science, WBS

Suzy Moat

Data Science Lab

Behavioural Science, WBS [email protected]

Quantifying human behaviour

using online data

Data Science Lab

The advantage of looking forward

1

Future Orientation Index 2010

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

Future Orientation Index 2010

Suzy Moat & Tobias Preis Based on Preis, Moat, Stanley and Bishop (2012)

Ratio of Google searches for “2011” to searches for “2009” during 2010 for 45 countries

more Google searches for “2009” more Google searches for “2011”

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

Future Orientation Index 2012

Suzy Moat & Tobias Preis

Based on Preis, Moat, Stanley and Bishop (2012)

Richer countries look forward

Time with Weekly Granularity

Search Volume

0

5

10

2008

2009

2010

0

5

A

B

“2008”

“2007”

“2009”

“2010”

“2011”

“2009”

Future-Orientation Index

GDP / Capita [10

4

USD]

1

2

3

4

0.0

0.5

1.0

1.5

2.0

    _{Preis, Moat,} Stanley & Bishop (2012) Featured by:

Photo: Perpetual Tourist

Anticipating market moves

Hypothetical strategy

week

t

Moat et al. (2013); Preis et al. (2013)

number of

week

t

t-1

t-2

t-3

Moat et al. (2013); Preis et al. (2013)

Hypothetical strategy

number of

Page views

decreased:

BUY

stock

in week t+1

week

t

t-1

t-2

t-3

Moat et al. (2013); Preis et al. (2013)

Hypothetical strategy

number of

Page views

decreased:

BUY

stock

in week t+1

Page views

increased:

SELL

stock

in week t+1

week

t

t-1

t-2

t-3

Moat et al. (2013); Preis et al. (2013)

Hypothetical strategy

number of

Wikipedia

: Dow Jones companies

Views data: signi

fi

cant di

ff

erence

Moat, Curme, Avakian, Kenett, Stanley & Preis (2013) Featured by:

Return [Std. Dev. of Random Strategies]

Density 0.0 0.2 0.4 0.6 −2 0 2 Wikipedia Views DJIA Companies Wikipedia Edits DJIA Companies Random Strategy

Wikipedia

: Financial topics

Moat, Curme, Avakian, Kenett, Stanley & Preis (2013) Featured by:

Views data: signi

fi

cant di

ff

erence

0.00 0.25 0.50 0.75 1.00 −2 0 2

Return [Std. Dev. of Random Strategies]

Density Wikipedia Views Financial Topics Wikipedia Edits Financial Topics Random Strategy

Wikipedia

: Actors and

fi

lmmakers?

Moat, Curme, Avakian, Kenett, Stanley & Preis (2013) Featured by: 0.0 0.1 0.2 0.3 0.4 −2 0 2

Return [Std. Dev. of Random Strategies]

Density

Wikipedia Views

Actors & Filmmakers

Random Strategy

Random strategy mean + 2 sds Random strategy mean + 1 sd return (random strategy sds)

0 1 2

-1

“debt”

“culture”

How

Google

keywords perform

Random strategy mean + 2 sds Random strategy mean + 1 sd return (random strategy sds)

0 1 2 -1

“debt”

“culture”

“stocks”

“credit”

“garden”

“train”

Preis, Moat & Stanley (2013)

# occurrences in FT

# hits on Google

Returns signi

fi

cantly

correlated with indicator

of

fi

nancial relevance

Financial relevance

Random strategy mean + 2 sds Random strategy mean + 1 sd return (random strategy sds)

0 1 2

-1

Preis, Moat & Stanley (2013)

debt money crisis internet technology money debt

Curme, Preis, Stanley & Moat (2014)

What is searched for

before falls?

55 groups of search

terms

Business and politics

most related

Curme, Preis,

Stanley & Moat (2014)

What is searched for before falls?

Cumulative Returns (%)

-100 0 100 200

Random Strategy Politics I Business

Photographers as sensors

Flickr

and tourist numbers

Seresinhe, Preis & Moat (under review)

Scenicness and wellbeing

Scenicness and wellbeing

A

D

Average percentage of greenspace 0 0.3 0.6 0.7 0.9 0.94 0.96 0.99

Average scenic rating

1 2.2 3.1 3.6 3.9 4.2 4.6 8 POOR HEALTH GREENSPACE SCENICNESS Average rates of poor health (SMR) 0 0.5 0.6 0.7 0.8 0.9 1.2 3.2 London Birmingham Manchester Newcastle Liverpool _Sheffield

B

C

0.00 0.25 0.50 0.75 1.00

All areas Urban Suburban Rural

Proba

blity of the model

gi

ven the data (AICw)

Model

Scenicness only Greenspace only

Scenicness and Greenspace

Scenicness and wellbeing

Seresinhe, Preis & Moat (under review)

A

D

Average percentage of greenspace

0 0.3 0.6 0.7 0.9 0.94 0.96 0.99

Average scenic rating

1 2.2 3.1 3.6 3.9 4.2 4.6 8 POOR HEALTH GREENSPACE SCENICNESS Average rates of poor health (SMR) 0 0.5 0.6 0.7 0.8 0.9 1.2 3.2 London Birmingham Manchester Newcastle

Liverpool _Sheffield

B

C

0.00 0.25 0.50 0.75 1.00

All areas Urban Suburban Rural

Proba

blity of the model

gi

ven the data (AICw)

Model

Scenicness only Greenspace only

Scenicness and wellbeing

Seresinhe, Preis & Moat (under review)

A

D

Average percentage of greenspace

0 0.3 0.6 0.7 0.9 0.94 0.96 0.99

Average scenic rating

1 2.2 3.1 3.6 3.9 4.2 4.6 8 POOR HEALTH GREENSPACE SCENICNESS Average rates of poor health (SMR) 0 0.5 0.6 0.7 0.8 0.9 1.2 3.2 London Birmingham Manchester Newcastle

Liverpool _Sheffield

B

C

0.00 0.25 0.50 0.75 1.00

All areas Urban Suburban Rural

Proba

blity of the model

gi

ven the data (AICw)

Model

Scenicness only Greenspace only

Scenicness and Greenspace

People report

better health in

more scenic

locations

Scenicness and wellbeing

A

D

Average percentage of greenspace 0 0.3 0.6 0.7 0.9 0.94 0.96 0.99

Average scenic rating

1 2.2 3.1 3.6 3.9 4.2 4.6 8 POOR HEALTH GREENSPACE SCENICNESS Average rates of poor health (SMR) 0 0.5 0.6 0.7 0.8 0.9 1.2 3.2 London Birmingham Manchester Newcastle

Liverpool _Sheffield

B

C

0.00 0.25 0.50 0.75 1.00

All areas Urban Suburban Rural

Proba

blity of the model

gi

ven the data (AICw)

Model

Scenicness only Greenspace only

Scenicness and Greenspace

Scenicness and wellbeing

A

D

Average percentage of greenspace 0 0.3 0.6 0.7 0.9 0.94 0.96 0.99

Average scenic rating

1 2.2 3.1 3.6 3.9 4.2 4.6 8 POOR HEALTH GREENSPACE SCENICNESS Average rates of poor health (SMR) 0 0.5 0.6 0.7 0.8 0.9 1.2 3.2 London Birmingham Manchester Newcastle

Liverpool _Sheffield

B

C

0.00 0.25 0.50 0.75 1.00

All areas Urban Suburban Rural

Proba

blity of the model

gi

ven the data (AICw)

Model

Scenicness only Greenspace only

Scenicness and Greenspace

A

D

Average percentage of greenspace

0 0.3 0.6 0.7 0.9 0.94 0.96 0.99

Average scenic rating

1 2.2 3.1 3.6 3.9 4.2 4.6 8 POOR HEALTH GREENSPACE SCENICNESS Average rates of poor health (SMR) 0 0.5 0.6 0.7 0.8 0.9 1.2 3.2 London Birmingham Manchester Newcastle

Liverpool _Sheffield

B

C

0.00 0.25 0.50 0.75 1.00

All areas Urban Suburban Rural

Proba

blity of the model

gi

ven the data (AICw)

Model

Scenicness only Greenspace only

Scenicness and Greenspace

Scenicness and wellbeing

A

D

Average percentage of greenspace 0 0.3 0.6 0.7 0.9 0.94 0.96 0.99

Average scenic rating

1 2.2 3.1 3.6 3.9 4.2 4.6 8

POOR HEALTH

GREENSPACE

SCENICNESS

Average rates of poor health (SMR) 0 0.5 0.6 0.7 0.8 0.9 1.2 3.2 London Birmingham Manchester Newcastle

Liverpool _Sheffield

B

C

0.00 0.25 0.50 0.75 1.00

All areas Urban Suburban Rural

Proba

blity of the model

gi

ven the data (AICw)

Model

Scenicness only Greenspace only

Scenicness and Greenspace

Measuring

fl

u with

Google

Ginsberg et al., Nature 368, 1012 (2009)

Flu data

Google Trends estimate

 

Butler, Nature 494, 155 (2013)

“The press reports may have triggered many fl

● ● ●●●●●●●●●_● ●● ●●●●●●●●● ●●●●●_{●●●●●●●}●●●●● ●●●● ●●● ● ●● ● ●● ●● ● ● ● ●● ● ● ● ● ● ● ● ●● ●●●●●●●_●●●● ●●●●●●●●●●● ●●●●● ●●●●●●● ●●_● ●●● ●●●●●● ● ● ●● ●● ●●●●●●●_●●●● ●●●●●●●●●●● ●●●●●● ●● ●● ● ● ● ● ● ●● ● ● ● ● ● ●●● ● ● ● ● ●●●●●●● ●●●●●_{●●●●●●●}●●●● 2 4 6 2010 2011 2012 2013 Time [Weeks] In fl uenza − Lik e Illness [%] Predicted Value Observed Value 80% Prediction Interval 95% Prediction Interval ● Training Period Out-of-Sample Nowcast

Preis & Moat (2014)

Flu estimate errors signi

fi

cantly reduced

To what extent can Internet data

help us measure and even predict

human behaviour?

[email protected]

@suzymoat

computer science statistics physics mathematics

crime science

finance health

economics