SNA An Introduction Jos Luis Molina joseluis molinauab
SNA: An Introduction José Luis Molina joseluis. molina@uab. cat http: //pagines. uab. cat/joseluismolina www. egolab. cat Màster en Direcció d'Empreses / Master in Management
What do we mean by “social networks”? Social Media (twitter, Facebook, Whatsapp…) A metaphor about complexity Market oriented: Social Networking Sites Adopted by Mass-Media. Convergenge of social and tecnological networks (Kleinberg 2008) … Castells, The Network Society(1999). (…) Emergent social structures Communities, social circles, quasi-groups (Boissevain 1968) … Transnational fields – diffusion of ideas, habits … Informal organizational chart. Industrial districts!
A long history. . .
Theoretical questions … Identification of the effect of emerging social structures in individual or collective behavior (rumors, Epstein 1969; strike, Kapferer 1972; political mobilization, Srinivas & Béteille 1964. . . ). Divergence between cognitive data and behavioral data. Relations between micro and macro social structures. Contagion - Influence, strong / weak ties. RDS. . . Intracultural variation of local ecological knowledge. Synchronicity of human activity (cities, regions, world). Other. . . (Adler-Lomnitz, Dunbar. . . ).
Basic definitions: centrality, cohesion, positions and structures
Whole networks– egonetworks, egocentric networks /personal networks A whole/complete network is a collection of nodes and ties taken from a specific field of interaction (a classroom, a organization, an internet community…).
Boundary specification (Laumann et al. 1983) Nominalistic Realistc Imposed by the researcher’s analytical interests (“the network of interlocking directorates in Spain”). Boundaries reported by the participants (“people in this neighboorhod”). Representative sampling is not possible (or by sampling 80%). Important/influential people can be easily identifed (indegree / nominations).
Ego-centric analysis: ego networks A “egonetwork” is the neigborhood of a given ego within a whole network. Every ego is a node of the whole network. The analysis focuses on the egonetwork level, including ego.
Ego-centric networks /personal networks A“personal network” is the set of people and ties of a given individual in different fields of interaction, along with the alter-alter ties. Ego is typically not represented. The analysis is conducted at the individual level. Representative sampling
Combination… Personal networks can be combined into a whole network.
Network data. . . • Name generators (cognitive data): questions that provide names of people in return. • Each name generator can generate a different network. • Digital activity logs (time, location, from, to, content … behavioral data). • Usually many observations and few ítems by each case (phone calls, Facebook likes, etc. ) • Others (letters, official registers, etc. )
Types of name generators… One name generator vs. multiple name generator Free list vs. fixed list. Free size / fixed number of nominations.
Network ties can be… Bynary (1, 0) or valued (e. g. 0 -5). Oriented or reciprocal
“Name interpreters” / attributes The list of descriptives of each alter nominated: Sociodemographics Ego-alter closeness. Geographic location. …
Types of matrices One-mode= same set of actors/nodes/vertices on rows and columns. Two-mode= one set of actors on rows and a different set of actors or events on columns. Also called Affiliation Matrices. It is possible to convert two-mode matrices to one-mode (rows * rows or columns * columns).
. . . Matrix one-mode. . . 1 2 3 4 5 6 7 8 9 10 11 12 13 14 AMG RVT GBB EG JSP JMF MOS CR PP GG JLM MAR FG DES 1 1 1 2 3 4 5 6 7 8 9 0 1 2 3 4 AMRVGBEGJSJMMOCRPPGGJLMAFGDE - - - - 0 0 0 1 1 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 0 1 0 0 1 1 0 0 1 0 0 0 1 0 0 0 0 1 1 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0
… and its sociogram. . .
Some basic concepts from graph theory … Path Geodesic Component Diameter …
Node level: centrality measures
Network level: density or av. degree, component, centre-perifery
Group-level Group detection following the cohesion strategy. Newman-Girvan Tab “Analysis”: grouping, clustering, modulartiu, girvan. Newman apply This operation creates a new attribute that can be seen in the Attribute Manager and gives a value to each group. This value can be used to differentiate / color / merge the different groups.
Collection of network data… Link List … Ego, alter, value: Tom Cruise, Nikole Kidman, 5 Tom Cruise, Penélope Cruz, 3 Nikole Kidman, Tom Cruise, 2 Nikole Kidman, Penélope Cruz, -5 o Nikole Kidman, Penélope Cruz, 0 (in case of absence, the dyad means a value of “ 1”). Attributes … Artist, sex, age, nationality, or degree of innovation, market share…
Import Link List + Atributes to visone. info Webstart or download Ejecutar el programa Descargar fichero CSV con los enlaces (en el escritorio). Descargar fichero CSV con los atributos (en el escritorio) Importar (file > open) en primer lugar el fichero links.
Attribute manager
Análisis a nivel de red Medidas de cohesión (densidad, reciprocidad, transitividad …). Componentes Estructura centro-periferia (…) Tab Analysis: task: indexing; class: network; index: network statistics, seleccionar densidad, average degree, conn. components, # nodes in max core. Apply Attribute Manager> Show/edit.
Some links. . . INSNA http: //www. insna. org/ Web Redes http: //www. redes-sociales. net Revista REDES http: //revistes. uab. cat/redes
- Slides: 27