Topics
• The explosion of social interactions• ICT and social interactions• Social Networks • Metrics• Data• Tools • Project ideas
2SSIIM, 2017/10/12
3SSIIM, 2017/10/12
SSIIM, 2017/10/12 4
SSIIM, 2017/10/12 5
SSIIM, 2017/10/12 6
Where are we now?
7SSIIM, 2017/10/12
SSIIM, 2017/10/12 8
SSIIM, 2017/10/12 9
Social media usage 2016
How teens communicate
SSIIM, 2017/10/12 10
SSIIM, 2017/10/12 11
Networks
• Networks are everywhere• Social, biological, financial, etc• Complex networks• Communities reveal properties of networks• Contagion• Controversies• Society!
12SSIIM, 2017/10/12
SSIIM, 2017/10/12 13
SSIIM, 2017/10/12 14
Social networks
Social interactions• Like, comment, share, cite • e-Commerce• Payments• Digital marketing• Political marketing• etc
15SSIIM, 2017/10/12
Where are we?
● Complex networks● Actors influencing and being influenced by
other actors● But humans are not software agents● Difficult to establish consensus● Intelligence highly needed● Maybe biology could inspire us...
SSIIM, 2017/10/12 16
SO?!
SSIIM, 2017/10/12 17
• Let's have a look at graphs and networks
SSIIM, 2017/10/12 18
Euler 1707 - 1783
SSIIM, 2017/10/12 19
Basics of graphs and networks
• G = (V, E)• O(G) = |V| order
• S(G) = |E| size
• A adjacency matrix
• Ki degree of vertex i
• Directed/undirected
SSIIM, 2017/10/12 20
Representation of networks
• Matrixes, graphs, edge lists, etc
A B C D EA 0 1 1 1 0B 1 0 1 0 1C 0 0 0 1 0D 0 1 1 0 0E 1 1 0 0 0
A BA CA DB AB CB EC DD BD CE AE B
SSIIM, 2017/10/12 21
• Equivalence relations– Reflexive, symmetric, transitive– Equivalence classes
• Order relations (partial, total or linear)– reflexive, anti-symmetrical, transitive– Hasse diagrams– x,y xRy yRx (total)
SSIIM, 2017/10/12 22
a bx taller than y
Be born in the same yearLive in the same street
Binary relations
SSIIM, 2017/10/12 23
Hasse diagram
m divides n
SSIIM, 2017/10/12 24
A OPO LIS MAD PAR LON BER A^2 OPO LIS MAD PAR LON BER A^3 OPO LIS MAD PAR LON BEROPO 0 1 1 1 0 0 OPO 3 0 1 1 2 2 OPO 2 7 6 6 1 1LIS 1 0 0 0 1 1 LIS 0 3 2 2 0 0 LIS 7 0 2 2 5 5MAD 1 0 0 1 0 1 MAD 1 2 3 1 1 0 MAD 6 2 2 5 3 5PAR 1 0 1 0 1 0 PAR 1 2 1 3 0 1 PAR 6 2 5 2 5 3LON 0 1 0 1 0 0 LON 2 0 1 0 2 1 LON 1 5 3 5 0 1BER 0 1 1 0 0 0 BER 2 0 0 1 1 2 BER 1 5 5 3 1 0
opo
lis
mad
parlon
par
parber
Composition of relations
• Usually not transitive (a likes b and b likes c but ...)
• “Equivalence” relations– No equivalence classes– But communities, clusters, etc
• “Order” relations (partial, total)– No Hasse diagrams– Rankings, proeminence indexes, etc
SSIIM, 2017/10/12 25
Real life relations
Global metrics
• Number of vertexes 5
• Number of edges 11
• Number of components 1
• Diameter 2
• Density 0.55
SSIIM, 2017/10/12 26
Centrality Measures
• Degree centrality– Edges per node (the more, the more important the node)
• Closeness centrality– How close the node is to every other node
• Betweenness centrality– How many shortest paths go through the edge node
• Bibliometric + Internet style (quality of edges)– PageRank, eigenvector
27SSIIM, 2017/10/12
Champions league Pagerank
SSIIM, 2017/10/12 28
Common Tasks
• Measuring “importance”– Centrality, prestige, influence (incoming links)
• Diffusion modeling– Epidemiological
• Clustering– Blockmodeling, Girvan-Newman, Chinese whisper
• Visualization/Privacy/etc.
29SSIIM, 2017/10/12
Community detection
• Communities and clusters are different • Network data is related to graph properties• Real world data is big
SSIIM, 2017/10/12 30
Modularity
• Compares number of edges with number of edges of a random network
• Maximize Q is NP-hard
SSIIM, 2017/10/12 31
jC,iCδij
ijPijAm21
Q
m2jkik
ijP
Dynamics
• Networks have a temporal dimension• Interactions – follow, like, share, mention,
retweet, hashtag, etc – occur in sequence• Network properties evolve in time
SSIIM, 2017/10/12 32
SSIIM, 2017/10/12 33
SSIIM, 2017/10/12 34
SSIIM, 2017/10/12 35
SSIIM, 2017/10/12 36
SSIIM, 2017/10/12 37
SSIIM, 2017/10/12 38
SSIIM, 2017/10/12 39
• Gephi• SocNetV• Netlogo
40SSIIM, 2017/10/12
Tools
Software Tools
• NetworkX• d3.js• API
SSIIM, 2017/10/12 41
Datasets
● Konect● SNAP● Gephi
SSIIM, 2017/10/12 42
Queen @ Spotify
SSIIM, 2017/10/12 44
Queen @ Spotify in Gephi
SSIIM, 2017/10/12 45
#barcelona @ Tumblr in Gephi
SSIIM, 2017/10/12 46
Ideas?
● Find and use APIs● Crawl Instagram● Hashtags co-occurrences (Twitter, Tumblr)● Detect fake/abnormal behaviours● Use your imagination!
SSIIM, 2017/10/12 47
Thank you!
SSIIM, 2017/10/12 48