Whoo.ly: Facilitating Information Seeking for Hyperlocal Communities Using Social Media

Presentation given by Yuheng Hu at CHI on paper about Whooly

HCI
1.0x

Whoo.ly: Facilitating Information Seeking for Hyperlocal Communities Using Social Media

By amh
Created 3 years ago

Duration 0:00:00
lesson view count 8
Presentation given by Yuheng Hu at CHI on paper about Whooly
Select the file type you wish to download
Slide Content
Tags: HCI
  1. Whoo.ly: Facilitating Information Seeking For Hyperlocal Communities Using Social Media

    Slide 1 - Whoo.ly: Facilitating Information Seeking For Hyperlocal Communities Using Social Media

    • Yuheng Hu
    • Shelly Farnham
    • Andrés Monroy-Hernández
  2. Sirens blaring in your neighborhood

    Slide 2 - Sirens blaring in your neighborhood

  3. Laura

    Slide 3 - Laura

    • This is Laura, she just moved into this
    • neighborhood
    • You are curious about the cause of commotion
  4. Check the local TV news

    Slide 4 - Check the local TV news

  5. Check the local newspaper

    Slide 5 - Check the local newspaper

  6. Search will probably be fruitless

    Slide 6 - Search will probably be fruitless

  7. Neighbors reporting relevant information on social media

    Slide 8 - Neighbors reporting relevant information on social media

  8. Social media:

    Slide 9 - Social media:

    • noisy
    • chaotic
    • hard to understand
  9. Whoo.ly reveals latent neighborhood information

    Slide 10 - Whoo.ly reveals latent neighborhood information

  10. Filter from Twitter:

    Slide 11 - Filter from Twitter:

    • Events
    • People
    • Places
  11. Active events 

    Slide 13 - Active events 

  12. Top topics

    Slide 14 - Top topics

  13. Active people

    Slide 15 - Active people

  14. Popular places

    Slide 16 - Popular places

  15. Event: topically related trending terms at a given time period

    Slide 17 - Event: topically related trending terms at a given time period

    • (Fung et al 2001)
  16. Identify

    Slide 18 - Identify

    • trending terms
    • Clustering
    • trending terms
    • Event Detection
  17. Exponential Moving Average

    Slide 19 - Exponential Moving Average

    • Moving Average Convergence/Divergence
    • Difference between two MACD
    • EMA
    • MACD
    • MACD Histogram
    • Identifying Trending terms
    • MACD histogram for the time interval is positive
    • A trending
    • term
  18. Clustering trending terms

    Slide 20 - Clustering trending terms

    • 1. Learn topics from trending terms
    • Latent Dirichlet Allocation
    • 2. Clustering the terms based on topical similarity
    • Shared nearest neighborhood clustering (SNN)
  19. Top Topics

    Slide 21 - Top Topics

    • normalized TF-IDF
    • Popular Places
    • Information Extraction
    • Active People
    • PageRank
    • Other Hyperlocal contents
  20. Evaluation

    Slide 22 - Evaluation

    • Semi structured interview
    • Neighborhood Information Seeking Task 
    • (n=13)
  21. Tasks: Whoo.ly VS. Twitter

    Slide 23 - Tasks: Whoo.ly VS. Twitter

    • recent events
    • local neighborhood reporters
    • neighborhood topics
    • potential neighborhood friends
  22. Using Whoo.ly is easier to complete tasks

    Slide 24 - Using Whoo.ly is easier to complete tasks

  23. Whoo.ly is more useful, easy to use, has better overview of neighborhoods, and a sense of connection to communities

    Slide 25 - Whoo.ly is more useful, easy to use, has better overview of neighborhoods, and a sense of connection to communities

  24. R1: Really liked it overall, definitely a lot easier to find stuff

    Slide 26 - R1: Really liked it overall, definitely a lot easier to find stuff

    • R2: Twitter isn’t set up for a community. Whool.y functions amazingly for this.
    • R3: Results. Mostly the furniture on craigslist. Need to filter out those, and be able to differentiate between the spammy ‘top users’ and the real top users.”
    • After finishing tasks
  25. Slide 27

    • Conclusion
    • Whoo.ly, a hyperlocal information gathering tool
    • Promising feedbacks from local residents
    • Thanks! Questions?
    • @hyheng
    • yuheng@asu.edu