Inteligência ColetivaInteligência Coletiva

Thiago H Silva

2

Collective Intelligence

Thiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG

?

3

Nowadays

Thiago H. Silva – DCC/UFMG

- Smartphones are extremelly popular

- Mobile data plans as well

Thiago H. Silva – DCC/UFMG

Many sensors

4

Nowadays

Thiago H. Silva – DCC/UFMG

- Smartphones are extremelly popular

- Mobile data plans as well

Thiago H. Silva – DCC/UFMG

Traffic monitoring Photo sharing Location sharing

Many sensors

Humans in the sensing process

Thiago H. Silva – DCC/UFMG

Collective Intelligence

Humans in the sensing process

Thiago H. Silva – DCC/UFMG

Collective Intelligence

Humans in the sensing process

Thiago H. Silva – DCC/UFMG

Collective Intelligence

Social sensors

Humans in the sensing processCollective Intelligence

Social sensors

Humans in the sensing process

Thiago H. Silva – DCC/UFMG

Collective Intelligence

Collective Intelligence

10

Collective Intelligence

Thiago H. Silva – DCC/UFMG

11

Collective Intelligence

Thiago H. Silva – DCC/UFMG

T1 T2 T3

Graph to study trajectories

Collective Intelligence

Thiago H. Silva – DCC/UFMG

T1 T2 T3

Can we those data to help large scale study of urban societies ?

13DCC/UFMGThiago H. Silva – DCC/UFMG

City Image

Revealing the City that We Cannot See.Thiago Silva, Pedro Vaz de Melo, Jussara Almeida, Juliana Salles, Antonio Loureiro. ACM Transactions on Internet Technology. 2014.

UFMG Microsoft Research

14

City Image

Belo Horizonte, Brazil New York, USA

15

City Image

Thiago H. Silva – DCC/UFMG

Belo Horizonte, Brazil New York, USA

School

Nightclub

Typical Monday

Bar

WorkRestaurant

Are transitions random?

City Image

Thiago H. Silva – DCC/UFMG

Belo Horizonte, Brazil New York, USA

School

Nightclub

Typical Monday

Bar

WorkRestaurant

There are more favorable transitions than others

City Image

DCC/UFMG

Transition graph: check-in accordingly to routines or local preference

Pizza place“Food”

User1Time 1

Thiago H. Silva – DCC/UFMG

18

City Image

DCC/UFMG

Transition graph: check-in accordingly to routines or local preference

Jazz bar“Nightlife”

User1Time 2

Pizza place“Food”

User1Time 1

Thiago H. Silva – DCC/UFMG

City Image

DCC/UFMG

Transition graph: check-in accordingly to routines or local preference

Food Nightlife

Transition user 1

Jazz bar“Nightlife”

User1Time 2

Pizza place“Food”

User1Time 1

Thiago H. Silva – DCC/UFMG

Transition graphNodes are main categories of places

20

City Image

DCC/UFMGThiago H. Silva – DCC/UFMG

21

City Image

Thiago H. Silva – DCC/UFMG

Cur

rent

loca

tion

Next location

Cur

rent

loca

tion

Next location

RejectionRejectionIndiferenceIndiference FavoringFavoring

22

City Image

Thiago H. Silva – DCC/UFMG

Cur

rent

loca

tion

Next location

Cur

rent

loca

tion

Next location

Invisible images

23

City Image

Thiago H. Silva – DCC/UFMG

Cur

rent

loca

tion

Next location

Cur

rent

loca

tion

Next location

24DCC/UFMG

Cultural Diferences

Thiago H. Silva – DCC/UFMG

You are What you Eat (and Drink): Identifying Cultural Boundaries by Analyzing Food & Drink Habits in Foursquare.Thiago Silva, Pedro Vaz de Melo, Jussara Almeida, Mirco Musolesi, Antonio Loureiro. ICWSM'14

UFMG UCL

25

Cultural diferences

DCC/UFMGThiago H. Silva – DCC/UFMG

Big challenge: find appropriate data to use

Traditional methods: surveys – Expensive– Hard to detect dynamic changes

We propose to extract this elements from social media

Cultural diferences

DCC/UFMGThiago H. Silva – DCC/UFMG

Eating & drinking habits are fundamental elements in a culture

27

Cultural diferences

DCC/UFMGThiago H. Silva – DCC/UFMG

Eating and drinking habits are fundamental elements in a culture

Map each user in

Cultural diferences

DCC/UFMGThiago H. Silva – DCC/UFMG

Eating and drinking habits are fundamental elements in a culture

Map each user in

Like questions in a Survey!

Cultural diferences

DCC/UFMGThiago H. Silva – DCC/UFMG

Eating and drinking habits are fundamental elements in a culture

Map each user in

represents if user likes Users answers

30

Cultural analysis of individuals

DCC/UFMGThiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG 30

31

Extraction of cultural signatures

DCC/UFMG

Spatial evaluation

Thiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG 31

Results for cities

32

Extraction of cultural signatures

DCC/UFMG

Spatial evaluation

Thiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG 32

Brazil

USA

Indonesia

Japan

Results for cities

33

Extraction of cultural signatures

DCC/UFMGThiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG

Temporal evaluation

33

Weekdays

34

Identifying cultural boundaries

DCC/UFMGThiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG 34

Each color/symbol indicates a cluster

Clustering areas inside cities

35

Identifying cultural boundaries

DCC/UFMGThiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG 35

London regions

Tokyo regions

NYC regions

Clustering areas inside cities

36

Identifying cultural boundaries

DCC/UFMGThiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG 36

Comparing with World Value Survey data

The similarities are striking!

You are What you Eat (and Drink): Identifying Cultural Boundaries by Analyzing Food & Drink Habits in Foursquare.Thiago Silva, Pedro Vaz de Melo, Jussara Almeida, Mirco Musolesi, Antonio Loureiro. ICWSM'14

37

Take away message

DCC/UFMGThiago H. Silva – DCC/UFMGThiago H. Silva – DCC/UFMG 37

There many opportunities to explore collective intelligence for innovation!

38

Thanks!

Thiago H. Silva – DCC/UFMG

Thanks!Obrigado!

www.dainf.ct.utfpr.edu.br/~thiagohs