isaia2012
TRANSCRIPT
Collective Background Extraction for Station Market Area by using Location Based Social Network
Kousuke KIKUCHI, PhD StudentWaseda University@kousukekikuchi
Oct 24 2012
Outline
IntroductionPrevious methods to study citySensoring on cityLBSN ~ another sensoring method on cityHypotheses and premisePurpose
Data Collection and VisualizationMarket Analysis MethodResultsConclusionFuture study
Introduction
Previous Methods
e,g. Oral History, Image of City
e,g. Person Trip, Mathematical Model
Introduction
space syntax the image of the city
quantitative qualitative
Introduction
Previous Methods
e,g. Oral History, Image of City
Problems:Limited number of examineesProbability for examinees to change their behavior
e,g. Person Trip, Mathematical Model
Problems:Ignorance of city dwellers’
diversity
How to sublate them?Sensoring the information with content!
Introduction
Sensoring Methods
Making sensor node
Utilization of “Big Data”
Introduction
Realtime Rome Project, MIT SENSEable City Lab
Introduction
Twitter:opens the data which can be accessed through APIcontains user name, time, text, location data (but only few
percent)has keywords related to daily lifecan be analyzed to evaluate not only place, but also user itself.
We think the premise to be suitable for twitter analysis.
Introduction
Premise
Our background never change Our background will appear in keywords of Social mediaCollection of our background will be the background in some area
City will emerge distinctive keywords feature.
Introduction
This thesis fabricates the system to extrapolate the market characteristics by assessing the user activities of Twitter and finds other potential function in a city
Data Collection and Visualization
Data Collection
Visualization
Market Analysis Method
Results
Results
sophisticated
Similar characteristi
cs
Conclusion
We assembled the programs to collect and analyze big data from twitter.We discovered users of Roppongi showed similar traits to Akihabara.This similarity may not be explained by the previous methods.
We can construct the methodology on clarifying the potential of city.
We pledge to work further to detect the distinctive feature in each stations by using more computational methods.
Contact
twitter: @kousukekikuchifacebook: Kousuke.Kikuchiacademia: KousukeKikuchi