Co-word lifeline

718 views

Published on

Published in: Education, Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
718
On SlideShare
0
From Embeds
0
Number of Embeds
3
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Co-word lifeline

  1. 1. Co-word Lifelines Noortje, Esther, Marieke, David & Carolin
  2. 2. Identifying key life signals Question: How lively is “climate action” on Twitter. Objective: Use the co-word machine as Issue Biographer. Changing co-words per intervalFocus on key hashtags Focus on intervals (Associational Profile)
  3. 3. Temporality & interval• Consideration: compare liveliness measures tofrequency of activity.• Intervals: problems.
  4. 4. Dataset• Twitter data “Climate Action”.• Co-word machine input: Tweets.• Focus on three intervals: 15Feb-14Mar, 15Mar-14Apr,15Apr-14May.• Objective: profile the co-word relations of keyhashtags.
  5. 5. Machine settings for the lifeline tracker • Normalised association strength. • Flattening co-word relations.
  6. 6. Machine settings for the lifeline tracker • Keyword profiling. • Determines changes in associational profile per interval. • Identifies degree of connectivity & change. • New words connecting, words disappearing & current connections.
  7. 7. Hashtag co-word network interval I
  8. 8. Hashtag co-word network interval II
  9. 9. Hashtag co-word network interval III
  10. 10. Hashtag profiling over time method 1. Detect key hasthags (av. weighted degree) per interval. 2. Determine URL profile for key hashtags per interval. 3. Determine co-word profile per hashtag per interval. 4. Determine overall variation per hashtags/for all hashtags.
  11. 11. Hashtag profiling•#tarsands •#jobs•#eu •#san•#cdnpoli •#intern•#agw •#job•#green •#cop18•#fqd •#cop17•#cndpoli •#climatechange•#politics •#energy•#unfccc •#globalwarming•#ceta •#environment•#health •#policy•#flooding •#losangeles •#nonprofit
  12. 12. Hashtag profiling #healthTop weighted degree hashtag #health: Only retweets
  13. 13. Hashtag actor profiling #green Interval I Interval II Interval III
  14. 14. Co-word machine as Issue BiographerCo-word machine input: Tweets, URLs, medium specific small text units,syntactic demarcation.Settings1. Determine temporal unit: interval or continuous.2. Identification of key words.- manually & measures (degree...).- across set or by interval3. Static vs. dynamic terms (signal words?)4. Determine associational profile per key word (int. or continuous?) by URLand keyword.
  15. 15. Keyword lifelineContinuous changes in associational profile as indicatorfor keyword liveliness.
  16. 16. Thank you.

×