***** To join INSNA, visit http://www.insna.org *****
we are happy to announce that on two occasions in the not so far future
we will be giving our workshop on ego-centred network analysis in R. Our
aim is to make the advanced techniques of (ego-centerered) network
analysis and visualisation available for users with all kinds of
The first workshop will be held at the Sunbelt 2016 in Newport Beach:
Tuesday April 5 6:30 - 9:30pm and
Wednesday April 6 8:00am – 11:00am
Sunbelt Workshop Program:
The second workshop will be at the EUSN 2016 in Paris, between the 14th
and 17th June (exact dates are to be disclosed).
Both workshops will entail a 6 hour program, packed with extensive
material and guided hands-on experience with R and R-Studio. Please
check out the links above to find more information (oder nochmal
komplettes Abstract anhängen?)
We’ll see you there.
Andreas and Till
(Sorry for double-posting!)
Simplifying ego-centered network analysis in R with egonetR
Workshop Length: 2 sessions (6 hours total)
Ego-centered network analysis is tricky because most often you have
n+many networks which you want to analyse simultaneously and performing
multivariate analyses, where network data and attributional data of ego
are combined. An ego-centered network is commonly known as the network
of a focal actor (ego), including the relationships of ego to alters and
the relationships between these alters. Analysis can be conducted on two
different levels: On the alter level the analysis focusses on the
qualities of the ego-alter-relationships e.g. provision of support or
contact frequency. The network level analysis focuses the structure and
composition of the ego-centered network. The most reasonable way to
analyse multiple ego-centered networks is to use R because it allows
both the calculation of various network measures for n+many networks
simultaneously (e.g. what is the density of all networks) and
multivariate analysis of network and attributional data (e.g. how does
density vary between men and women). The workshop focuses the analysis
of ego-centered networks using data from current research projects. We
introduce the R-package “egonetR” which simplifies the import and
manipulation of ego, alter and network-level data in different formats
and allows to calculate a range of network measures. Using “egonetR”
users will be enabled to conduct ego-centered network analysis with very
basic R programming knowledge.
After a short introduction into data management and data import to R we
calculate network measures on alter level (e.g. multiplexity, homophily)
and network level (size, density, EI-Index, diversity, components,
proportions of ties with specific attributes). Afterwards we go on with
multivariate analyses, both on alter and the network level. For both
levels we exemplify explorative and hypothesis testing procedures using
additional packages in R (e.g. cluster, lme4, FactoMineR): On the
network level we focus cluster analysis and standard regression. With
cluster analysis we present a way to typologize ego-networks along
network-level information. In standard regression analysis, network
measures can be used as dependent or independent variables, just as any
other characteristic of ego e.g. to test whether network size differs by
gender or age. On the level of the analysis of ego-alter dyads we
demonstrate multivariate correspondance analysis and give a short
introduction to multilevel analysis. Multivariate correspondence
analysis is used to explore the dependencies between categorial
alter-level variables using visualisations. Multilevel analysis is used
to treat with the nested nature of alter level data when dependent
variables lies on the alter level.
Major parts of the workshop will be “hands-on” utilizing R (R-Studio). A
short intro into the basics of R will be given in the beginning while
prior knowledge in R is preferable.
For more in depth introductions to R it is recommended to visit the
workshop „Introduction to ego-network analysis with R“ by Raffaele
Vacca, where data collection, data management and data transformation
between different levels of data are discussed in detail and Michal
Bojanowski's "Introduction to R and Social Network Analysis with igraph"
for a general introduction to (whole) network analysis with igraph in R.
Further information will be available on (http://rego.tillt.net).
Technology Needs: Laptops with internet access. Installation of R
(http://cran.r-project.org/) and R-Studio (www.rstudio.com) before the
workshop is required.
Instructors: Till Krenz and Andreas Herz
Intstitutions: Till Krenz (Otto-von-Guericke University Magdeburg),
Andreas Herz (University of Hildesheim)
Email: [log in to unmask]
Till Krenz, M.A.
Institut für Soziologie
Tel. 0391 - 67 56 596
Fax. 0391 - 67 16 533
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.