Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection,...

49
http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan Huang, Alina Lungeanu, York Yao, Noshir Contractor [email protected] Science of Networks in Communities (SONIC) Northwestern University Cyberinfrastructure for Inquiring Knowledge Networks on the Web Developed by York Yao, Yun Huang, and Jinling Li in the SONIC research group at Northwestern University. C-IKNOW evolves from I-KNOW, previous contributors include Dan Zink, Mike Chan, Peter Taylor, Dana Serb, Ryan Kanno, Mike Armstrong, Shyam Kurup, Emily Wang, Sean Mason, Jered Wierzbicki, Jeff Tamer, Hank Green, Steven Harper, Nat Bulkley, Andy Don. Supported primarily by grants from the National Science Foundation (NSF) (Award # 904356, # 838564, # 836262, #753047, #9980109) & and the National Institutes of Health (NIH).

Transcript of Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection,...

Page 1: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and RecommendationMeikuan Huang, Alina Lungeanu, York Yao, Noshir [email protected] of Networks in Communities (SONIC)Northwestern University

Cyberinfrastructure for Inquiring Knowledge Networks on the Web

Developed by York Yao, Yun Huang, and Jinling Li in the SONIC research group at Northwestern University. C-IKNOW evolves from I-KNOW, previous contributors include Dan Zink, Mike Chan, Peter Taylor, Dana Serb, Ryan Kanno, Mike Armstrong, Shyam Kurup, Emily Wang, Sean Mason, Jered Wierzbicki, Jeff Tamer, Hank Green, Steven Harper, Nat Bulkley, Andy Don.

Supported primarily by grants from the National Science Foundation (NSF) (Award # 904356, # 838564, # 836262, #753047, #9980109) & and the National Institutes of Health (NIH).

Page 2: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

C-IKNOW (Cyberinfrastructure of Inquiring Knowledge Networks on the Web): OverviewNoshir [email protected]

Science of Networks in Communities (SONIC)Northwestern University

Sunbelt Workshop, 12pm-3pm, Feb. 8, 2011, St. Pete Beach, Florida

Page 3: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

3

Why C-IKNOW

• Web-based platform integrates multiple tools for network analysis research• C-IKNOW surveys and direct data upload• Built in network analytics and visualization tool• Recommendation system: resource finder and

team assembly• Give value to users by allowing access to

visualization and recommendation tools

Page 4: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

4

Freeman’s EIES Dataset • EIES: Electronic Information Exchange System• Experiment published in 1980 on computer mediated

communication, early email• Data: 32 nodes

• 2 Attributes• Discipline: Questionnaire• Citations: Social Science Citation Index

• 3 Relations• All messages sent: Server data• Acquaintance at Time 1: January, 1978• Acquaintance at Time 2: September, 1978

Page 5: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

5

Sunbelt Workshop

• http://iknow.northwestern.edu/sunbelt2011.html • Today’s agenda:Time Topic Tutorial

12-1pm Data collection & survey design 1.1, 1.2, 1.3

1-1:50pm Visualization & analytics 2.1, 2.2

1:50-1:55pm Short break

1:55-2:30pm Recommender 3.1, 3.2, 3.3

2:35-3pm Q&A

Page 6: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

C-IKNOW (1): Data Collection: Web-Administered Survey & Direct Data UploadAlina LungeanuNoshir [email protected]

Science of Networks in Communities (SONIC)Northwestern University

Page 7: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

7

Overview of C-IKNOW Data Collection1. Survey Monkey for network

data2. Distributed researchers,

distributed participants3. Mix ‘n’ mash your data4. Allows imports and exports

to a wide variety of network tools and formats

Page 8: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

8

1. C-IKNOW’s 19 Question Types

Node AttributesChoice

Standard

Multi

Perceived

Rating

Standard

Multi

Perceived

Text

Standard

Long

Continuous

RelationsChoice

Standard

Multi

Perceived

Rating

Standard

Multi

Perceived

ContinuousMulti Relational Choice

Perceived Relational Rating(CSS )

Page 9: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

9

C-IKNOW step-by-step Survey Design

Page 10: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

10

2. The Web-Based Advantage

• Facilitate distributed collaboration among network researchers

• Improve distributed data collection across dispersed populations

Page 11: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

11

Built in Survey Invitations

• Automatically distribute usernames & passwords

Page 12: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

12

Survey Progress Reports

• Real-time reports monitor respondent progress

5/14

Page 13: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

13

C-IKNOW step-by-step User Administration

Page 14: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

14

3. Mix ‘N’ Mash Your Data• Supports digitally harvested

data & survey data

• Each question type has a unique data template

• Import/export DL, graphML, .txt, linklist, and other files

• Choose to import/export all or single questions

Page 15: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

15

C-IKNOW step-by-step Import / Export Capabilities

Page 16: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

16

4. Designed to Leverage Capabilities of Other Software Tools

C-IKNOW

Network Analysis Software Tools:

UCINET, Pajek, SIENA, STATNET,

PNet, NodeXL, etc.

3rd Party Data:Web of Science, Hyperlink data, (LexiURL, Voson)

Computer Log data, etc.

Excel

C-IKNOW Survey Data

Page 17: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

C-IKNOW (2): Web-based Network Data Visualization & AnalysisMeikuan HuangNoshir [email protected]

Science of Networks in Communities (SONIC)Northwestern University

Page 18: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

18

CIKNOW Visual Analytics: Highlights

• State-of-the-art interactive visualization

• A web-based platform for collaborative exploration • Empowers both researchers

and participants• Offers a central data

repository• Encourages collaboration

Page 19: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

19

Network Data Analytics—1. Visual Analytics• Explore the data: maneuver the network

• Layout, cluster, size, shape, weight• Detailed data

• Double click any edge or node for data details

• Save visualizations• Return to key findings

• Demo

Page 20: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

20

Network Data Analytics—2. Metrics

• Get all descriptive SNA stats: Just a click away• View & Export Analytics; • View & Export Charts;

• Derive new network relationships• Analyze• Visualize

• Even more: Export data files• Analyze with PNet• Analyze with UCINET• Analyze with SIENA

Pearson Correlation: 0.790***

Rate parameter .8972*(SD = .2402) Reciprocity: 3.7060* (SD = .7531)

Time 2

Time2:

Time1 & Time2:

Time1 & Time2:Estimate SD t-value

arc -4.191779* .27361 -.02288reciprocity 4.635971* .60276 .05773

Time1 & Time2:

Page 21: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

21

Page 22: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

C-IKNOW (3): Web-based Network Recommender SystemNoshir [email protected] LungeanuMeikuan Huang

Science of Networks in Communities (SONIC)Northwestern University

Page 23: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

23

C-IKNOW Recommendation System

Survey + External Data

RECOMMENDATION ALGORITHMS

1: Resource Discovery

Attributes Relations

2: Team Assembly

RECOMMENDATION ALGORITHMS

Page 24: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

24

Resource Discovery: Topics

1. The Basics2. Example with C-IKNOW3. System Setup4. Additional Capabilities

Page 25: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

25

Resource Discovery 1: The Basics

How am I connected to a particular person (e.g. Claude Fischer)?

How can I find individuals with a particular attribute?

Page 26: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

26

Resource Discovery 2:

A Statistician in Search of an Expert Sociologist

Page 27: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

27

Resource Discovery 3: System Setup• Select methods for part A and part B:

• Shortest Path• Similarity

• Select relevant relationsBA

Paul Holland Claude Fischer

Page 28: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

28

Resource Discovery 3:

System Setup

Page 29: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

29

Resource Discovery 4: Additional Capabilities• Multiple Node Types

• User, Organization, Keyword, etc.• Scores• Visualization

Page 30: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

30

Team Assembly: The Basics

Partition to give groups…

Max Density Similarity on one attribute Differences on a 2nd attribute

Page 31: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

31

Team Assembly:

An Example: Assembling Co-located

Interdisciplinary Teams of Experts

Page 32: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

32

Team Assembly: Algorithm Input

• Survey data (attribute and relational)• Team Specs• Configuration Data

• ω1 + ω2 + ω3 = 1• Diversity attribute priority levels (β1, β2, .. βn)

Page 33: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

33

Team Assembly: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices (specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:Until all individuals with that attribute are assigned, add an individual to a team by given method

4. For undersized teams: Add individuals who contribute least to existing teams that are not undersized

Page 34: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, & RecommendationNoshir [email protected] HuangAlina LungeanuYork Yao

Science of Networks in Communities (SONIC)Northwestern University

Page 35: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, & RecommendationDeveloped by York Yao, Yun Huang, and Jinling Li in the Science of Networks in Communities (SONIC) research group at Northwestern University. C-IKNOW evolves from I-KNOW, previous contributors include Dan Zink, Mike Chan, Peter Taylor, Dana Serb, Ryan Kanno, Mike Armstrong, Shyam Kurup, Emily Wang, Sean Mason, Jered Wierzbicki, Jeff Tamer, Hank Green, Steven Harper, Nat Bulkley, Andy Don.

©2009 Noshir Contractor, Science of Networks in Communities (SONIC), Northwestern University, Evanston IL

Page 36: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

36

Team Assembly 4: Evaluation Function = ω1*(β1*I{a1}+β2*I{a2}+..+βn*I{an}) (Diversity)

+ ω2*(1/m)*(Σ {j=1..m} pj*pj) (Similarity)

+ ω3*(network density) (Density)

Such that: I{ai} is 1 if attribute ai is present, and 0 otherwise

n is the number of options within the diversity attribute

pj is the percent of individuals who possess aj

m is the number of options within the similarity attribute

Page 37: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

37

Team Assembly 5: Assignment Method

1. Calculate score for each

partially formed team

2. Calculate score when individual is

added to each team

4. Tie: assign to lowest-

scoring team

3. Assign individual to team with

largest increase

Initial Score

0.08 0.11

Team 1 Team 2

Team 1!

Score with Individual

0.10 0.13

Change in Score

0.02 0.02

Page 38: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

38

Team Assembly 6: Algorithm

1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team

by given method

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 39: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

39

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices (specified

βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team

by given method

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

Page 40: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

40

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 41: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

41

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 42: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

42

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 43: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

43

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 44: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

44

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 45: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

45

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 46: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

46

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 47: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

47

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an individual to a team by given

method (highest score)

4. For undersized teams: Add individuals who

contribute least to existing teams that are not

undersized

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 48: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

48

Team Assembly 6: Algorithm1. Create t empty teams

2. Order critical diversity attribute choices

(specified βi’s or default)

3. FOR EACH DIVERSITY OPTION:

Until all individuals with that attribute are assigned, add an

individual to a team by given method

4. For undersized teams: Add individuals who contribute least to

existing teams that are not undersized

Math/Statistics (β1):

Anthropology (β2):

Sociology (β3):

Other (β4):

TEAM 1 TEAM 2

TEAM 3 TEAM 4

Page 49: Http://iknow.northwestern.edu/ C-IKNOW: An Integrated Web-based System for Network Data Collection, Analysis, Visualization, and Recommendation Meikuan.

http://iknow.northwestern.edu/http://iknow.northwestern.edu/

49

Team Assembly 7: Summary

Greedy method

Perform 10 iterations (default)

• Randomness from ties

Identify the best solution for each objective:

• Maximize the minimum group score• Minimize the variance of all group scores

Lenovo User
Vey nice design. But try miniizing the colors here.