Social data mining on smartphones
Upcoming SlideShare
Loading in...5
×
 

Social data mining on smartphones

on

  • 747 views

 

Statistics

Views

Total Views
747
Views on SlideShare
747
Embed Views
0

Actions

Likes
0
Downloads
10
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

Social data mining on smartphones Social data mining on smartphones Presentation Transcript

  • ETH D-ITET Eidgenössische Technische Hochschule Zürich Computer Engineering and Networks Laboratory Swiss Federal Institute of Technology Zurich Communication Systems Group Social Data Mining on Smartphones Semester Thesis Ajita Gupta Advisor: Sacha Trifunovic Supervisor: Prof. Dr. Bernhard PlattnerThursday, 7th July 2011 Ajita Gupta
  • Outline 1. Introduction 2. Related Work 3. Design 4. Implementation 5. Application Deployment 6. Evaluation Results 7. ConclusionThursday, 7th July 2011 Ajita Gupta
  • Introduction Motivation  Endless number of possibilities to connect with people  Applications must exploit heterogeneity and behavior patterns  Social Data Mining Systems record and analyze user activities  User satisfaction is improvedIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Introduction Problem Statement 1. Collect social contacts 2. Analyze and correlate social interactions 3. Classify communication patterns  Android Application: SocialMineIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Related Work Stumbl  Aim Define patterns between mobility, social connections and communication of users  Methodology Facebook Application  Difference to our approach + We capture more details and interaction types + Negligible user effort - Limited set of usersIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Related Work Device Analyzer  Aim Improve future smart phones, extract patterns and trends  Methodology Android Application  Difference to our approach + Focus on more social data (mainly from Facebook) - No incentive (e.g. live statistics)Introduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Design Social Data Mining SystemIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Implementation Extraction Modules - Contact Names - Caller - Sender - Timestamps - General Profile - Contact Numbers - Callee - Recipient - Friends - Email Addresses - Timestamps - Timestamps - Latitude - Messages - Postal Addresses - Call Duration - SMS Length - Longitude - Wallposts - Organization - Hobbies - Notes - GSM Cell - Pokes - Websites Location (Cell-ID, - Likes - Instant LAC) - Groups Messaging ID’s - BSSID’s - Events - Signal Strength of AP’sIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Deployment Specifications  Recruitment Invitation E-Mail  9 candidates  Duration Friday, 10th June – Friday, 17th June, 2011  Evaluation Feedback E-Mail  User SurveyIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Evaluation 1) Contact Graph  Number of Social Ties Contacts list: 308 Facebook friends: 208 (> 130: Facebook Statistics) Overlappings: 61  Conclusion High discrepancy between both contact setsIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Evaluation 2) Interaction Graph  Correlations between the contact sets of individual communication mediums  Conclusion - Helpful to classify contactsIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Evaluation 3) Communication Patterns Graph  Phone - Call Count: 6.7 calls/day/user - Call Duration: 1m 54s/call  SMS - SMS Count: 3 SMS/day/user - SMS Length: 73 chars/SMS  Location - User’s Daily Time Schedule * Workplace: 6h 46min * Home: 13h 31min * Meals: 1h 24 min * Travelling: 2h 17 min - Interaction Partners: 5.78 colleagues/day/user - Interaction Duration: 2h 21m/day/userIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Evaluation 3) Communication Patterns Graph  Facebook (most recent) - Message Threads: 13 - Wallposts: 22 - Likes: 8 - Pokes: 0.8 (from and to) - Groups: 26 - Events: 0.4  Conclusion - Characterize users (e.g environment) - Determine the tie strength of relationships - Derive user preferencesIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Conclusion Summary  Basic Social Data Mining System (built upon SocialMine)  Collection of social metadata with five different data extraction modules  Inspection of three different social dimensionsIntroduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Conclusion Outlook  Pilot stage of an iterative design process of an efficient Social Data Mining System  Areas for improvement 1. Add User Control 2. Add Data Anonymization Technique 3. Include User Attraction  Next Step: Classification and Profiling of User Behavior (SocialMine in the background + User-focused activity in the foreground + large-scale deployment)Introduction Related Work Design Implementation Deployment Evaluation Conclusion
  • Questions?Thursday, 7th July 2011 Ajita Gupta