Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

From Migration Corridors to Clusters: The Value of Google+ Data for Migration Studies

No description
by

Johnnatan Messias

on 22 August 2016

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of From Migration Corridors to Clusters: The Value of Google+ Data for Migration Studies

From Migration Corridors to Clusters:
The Value of Google+ Data for Migration Studies
johnnatan@dcc.ufmg.br
Asonam 2016
San Francisco - USA
Introduction
Data filtering
Expected Migration Flows
From Migration Corridors to Clusters:
The Value of Google+ Data for Migration Studies
More than 2 countries people have lived in
Google+
300 mi activity users
"Places Lived"

Provide a dataset for migration studies
Identify and study cases higher or lower than expected
Identify a set of features related to migration clusters
Scarina
Google+ Dataset
Google+ Dataset
All Google+ data until June 2012
160,304,954 users
Geolocation
Places Lived field (free text)
São Paulo == Sampa e EUA == USA
Belo Horizonte, Minas Gerais, Brazil
22,578,898 (14%) geo-mapped location
Occupation and Occupation Groups
Relationship status
Gender
Conclusion
Conclusion
First kind of study about migration clusters

Migration Clusters cannot be identified using information about corridors alone.

Limitations

No single dataset is enough to study international migration

International Migration Dataset is freely available

Explaining Deviance from Expectation
What is the importance in
International Migration studies?
Proposal
Existing traditional sources vs Web data
Multidisciplinary study
Johnnatan Messias, Fabrício Benevenuto, Ingmar Weber,
Emilio Zagheni
http://johnnatan.me
Our dataset enables us to identify clusters of countries
Frequency of each cluster

Our goal is to identify and study cases where observed counts of people who have lived in 3 countries are higher or lower than expected
Who are the migrant users?
1,958,656 migrants
1,565,803 have lived in 2 countries
271,142 have lived in 3 countries
69,129 have lived in 4 countries
52,582 at least 5 countries
192 countries (at least 1,000 people)

More features have been included

Classes of Clusters
Corridors vs Clusters
{BR, HU, US}
{IN, UK}, {IN, US}, {UK, US} -> {IN, UK, US}
Expected Migration Flows
To discover such patterns we used 4 formulas

(i) Ranking 1 ~ freqAB + freqAC + freqBC
(ii) Ranking 2 ~ freqAB * freqAC * freqBC (
gravity models
)
(iii) Ranking 3 ~ min(freqAB, freqAC, freqBC)
(iv) Ranking 4 ~ min(freqAB, freqAC, freqBC) * mean(freqAB, freqAC, freqBC)
As expected
: The center 20% of the cluster with expected and actual ranks approximately equal
Higher than expected
: Clusters that appear in the top 20% on the positive side
Lower than expected
: Clusters that appear in the top 20% on the negative side.
Thanks to Gabriel Magno for sharing his data collection
Features
Migration patterns depend on a multide of factors
Common Civilization
Common Colonial Link
Visa Requirement
Geographic Distance (Max, Avg, Min)
GDP: Gross Domestic Product (Max, Avg, Min)
Feature Importance
Full transcript