2. KBDeX is
• Knowledge Building Discourse eXplorer
• To support Social Network Analysis for KB discourse
• To support qualitative analysis (not a big data /
quantitative analysis approach)
• Strong for contents level, and temporal analysis
• Free for download, Work with both KF4, and 5
3. CHOISE
SNA SUPPORT SOFTWARE
KBDeX
Temporal
Network Analysis
Static
Network Analysis
Content
Network
Human
Network
gephi condor3
4. KBDeX:
Knowledge Building Discourse Explorer
network of
discourse units
Presented in
COINs 2010, COINs 2011
network of
students
network of
words
Input:
・Discourse Data(CSV format)
・Selected Words
Output:
・Three networks
・Temporal graph for metrics
6. Network Creation
Bipartite Graph of notes and words
A network structure of words based on the co-occurrence
within same notes
1 2 3 4
α β γ δ ε
Word
Note
α
β
γ δ
ε
7. Network Creation
Bipartite Graph of notes and words
A network structure of words based on the co-occurrence
within same notes
Another network structure of notes based on the co-occurrence
of words
1 2 3 4
α β γ δ ε
単語
1 2
3 4
word
Note
α
β
γ δ
ε
8. Network Creation
Bipartite Graph of notes and words
A network structure of words based on the co-occurrence
within same notes
Another network structure of notes based on the co-occurrence
of words
1 2 3 4
α β γ δ ε
単語
1 2
3 4
word
Note
α
β
γ δ
ε
9. Network Creation
A network structure of the students,
1 2 3 4
α β γ δ ε
単語
1 2
3 4
word
Note
α
β
γ δ
ε
Students
10. Network Creation
A network structure of the students, is based on sharing
the words
1 2 3 4
α β γ δ ε
単語
1 2
3 4
word
Note
α
β
γ δ
ε
Students
15. DEMO
• JICS Grade 4, 22 students, 3 months, 380 notes
• Science class, How light travels
• 123 manually selected words (from Grade1-8
curriculum standards — thanks to Jianwei)
• Decentralized Student network
• Rotating Leadership: 20 / 22 students took a important
position in betweenness centrality. Potential to be an
indicator of collective cognitive responsibility.
We are proposing the took KBDeX, which is designed to analyze discourses for researchers and students.
three networks and their temporary evolution over time.
We use a simple strategy to create the network from discourse.
First, we make a Bipartite Graph of notes and words.
Second, we make a one-mode projection of the network for words,
A network structure of words represents the co-occurrence within same notes.
Third, we make another one-mode projection for notes to the opposite side.
A network structure of notes represents the co-occurrence of words.
Third, we make another one-mode projection for notes to the opposite side.
A network structure of notes represents the co-occurrence of words.