Spatio-Temporal Dynamics of Geo-tagged Tweets

1

Sptio-Temporal Dynamics of Geo-tagged Tweets in

West Lafayette, IN

Yue Li

li1050@purdue.edu

2

Introduction

• Twitter

− The most popular micro-blogging site

− Tweets with longitude and latitude

− A gold mine for scholars in linguistics, sociology, economics, health, and psychology (Ghosh & Guha, 2013)

• West Lafayette, IN

• Most densely populated city in IN

• Home of Purdue University

3

Methodology

• Collect 4160 geo-tagged tweets using the Twitter Streaming API from April 11, 2013 to April 18

• Compare the spatial distribution of geo-tagged tweets on weekdays with those at the weekend

− Point Density tool in ArcGIS 10.1

− Clustering in Esri Maps for Excel

• Analyze the tweets on an hourly basis

4

Results

• Geo-tagged Tweet Clusters on Weekdays

5

Results

• Geo-tagged Tweet Clusters on Weekend

6

Results

7

Results

8

Results

0

50

100

150

200

250

300

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

NU

MB

ER

OF

TW

EE

TS

TIME

COUNT OF GEO-TAGGED TWEETS BY HOUR

weekday weekend

9

Results

• Geo-tagged Tweet Clusters from 11AM to 12PM

10

Results

• Geo-tagged Tweet Clusters from 20PM to 21PM

11

Conclusion

• Analyze the sptio-temporal pattern of geo-tagged tweets to discover the human mobility pattern hidden behind

• Proves the feasibility of using geo-tagged tweets, in local market research, market promotions, human mobility analysis, and even education regulation in a “college town” such as West Lafayette

• Future work

− Semantic analysis, topic modeling, and content analysis, aiming to track the spread of ideas and thoughts in local area

− Framework of extracting spatio-temporal social patterns from geo-tagged tweets in a city scale to help social researchers, demographic surveyors, market researchers, advertisers, and policy makers

12

References

• Ghosh, D., & Guha, R. (2013). What are we ‘tweeting’aboutobesity? Mapping tweets with topic modeling and Geographic Information System. Cartography and Geographic Information Science, 40(2), 90-102.

• Google Earth