Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR...
Transcript of Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR...
![Page 1: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/1.jpg)
SALESDECK2016v2
Social Data Science
DR PAUL SIEGEL/ DATA SCIENTIST
![Page 2: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/2.jpg)
SALESDECK2016v2
You have a good sentiment model. Now what?
● Identify influencers
● Construct audiences
● Activate advocates Key model: social networks
![Page 3: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/3.jpg)
SALESDECK2016v2
NOW YOU KNOW | #NYKCONF
BRANDWATCH.COM
![Page 4: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/4.jpg)
SALESDECK2016v2
Social Network Analysis
1. Build network model from data
2. Community detection
3. Influence modelling / optimization
![Page 5: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/5.jpg)
SALESDECK2016v2
Network Models
![Page 6: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/6.jpg)
SALESDECK2016v2
Social graph: nodes are people, edges are: ● Interactions
● Followership
● Latent influence
![Page 7: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/7.jpg)
SALESDECK2016v2
Coupled Hidden Markov Models
![Page 8: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/8.jpg)
SALESDECK2016v2
Community Detection
![Page 9: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/9.jpg)
SALESDECK2016v2
Isoperimetric Problem
![Page 10: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/10.jpg)
SALESDECK2016v2
Algorithms: ● Fiedler vector
● Louvain
● Short random walks
![Page 11: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/11.jpg)
SALESDECK2016v2
Short Random Walks
![Page 12: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/12.jpg)
SALESDECK2016v2
Influence
![Page 13: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/13.jpg)
SALESDECK2016v2
● Linear threshold model
● Independent cascade model
● Centrality heuristics (degree, eigenvector, PageRank, etc.)
![Page 14: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/14.jpg)
SALESDECK2016v2
Tech
![Page 15: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/15.jpg)
SALESDECK2016v2
● Networkx (Python)
● Graphx / GraphFrames (Spark)
● Graph databases?
![Page 16: Social Data Science2017.sentimentsymposium.com/.../2017/...PaulSiegel.pdf · Social Data Science DR PAUL SIEGEL/ DATA SCIENTIST ... Construct audiences Activate advocates Key model:](https://reader033.fdocuments.us/reader033/viewer/2022050407/5f846fa1ccc6b873f63d73e2/html5/thumbnails/16.jpg)
SALESDECK2016v2
References
● Everyone’s an Influencer: Quantifying Influence on Twitter by Watts et al ● Modeling Temporal Activity Patterns in Dynamic Social Networks by
Vasanthan Raghavan et al ● Four proofs for the Cheeger inequality and graph partition algorithms by
Chung ● Maximizing the Spread of Influence through a Social Network by Kempe et
al ● networkx.github.io ● spark.apache.org/graphx