Mobie Big Data Analysis for User's Behavior with IPython

Yong-il Lee, Wayne Jo, Youngdeok Kim

Audience level:


There are many statistics package, for instance R, MATLAB, SPSS and etc. Python also has various libraries for statistic analysis. IPython is a very attractive tool to do these analyses easily and briefly. If we combine other frameworks, libraries with python, we can visualize the analyzed result and apply to services in real time. Our main target data is personal information of mobile devices.


Mobile devices, especially smart phone, have enormous data of each user's already. So our goal is understanding user's behavior based on personal information on mobile devices. We want to detect user's state, emotion, preference from the information. SMS, call record, mail, app history imply user's intention and preference. We define this work as mobile big data analysis. It includes visualizing the data and sharing each person comfortably. Whole analysis process will be done python platform. Core functions are 1) real time analysis for user's state, 2) life-cycle period data gathering and auto-backup, 3) online visualization and sharing. Some parts needs Java and Object C++ programming to implement mobile apps. But main analyses are made by python packages. Our target data are like below. Stage 1 Call log, voice recording, SMS, Contact, Location(GPS), Photo Stage 2 App install/usage history, Web surfing pattern, Search record Stage 3 SNS(Twitter, facebook), Online BBS/community, mail, schedule