Personal networks (tutorial)
EgoNet is a software to conduct interviews in which the personal networks of respondents are collected. This tutorial explains (1) how to load data collected with EgoNet into visone and (2) how to cluster, aggregate, and visualize collections of personal networks using the methodology proposed in: Ulrik Brandes, Juergen Lerner, Miranda J. Lubbers, Chris McCarty, and Jose Luis Molina "Visual Statistics for Collections of Clustered Graphs". Proc. IEEE Pacific Visualization Symp. (PacificVis'08), 2008 (link to pdf).
- 1 An exemplary dataset
- 2 The EgoNet2GraphML software
- 3 Converting EgoNet interviews into GraphML files
- 4 Visual analysis of personal networks on the individual level
- 5 Class-level analysis of personal networks
- 6 Tendency and dispersion in collections of personal networks
An exemplary dataset
The data we are going to use for illustration in this tutorial have been collected within a study analyzing personal networks of immigrants in Barcelona. The study has been conducted by EgoLab and funded by the Fundació ACSAR pel Comissionat per Immigració i Diàleg Intercultural de l’Ajuntament de Barcelona. For more on this study and its outcome see the book (in Catalan): Jose Luis Molina and Fabien Pelissier (eds.) (2010). Les xarxes socials de sikhs, xinesos i filipins a Barcelona. Barcelona: Fundació ACSAR. (Also see the following link.)
The data consists of 70 EgoNet interviews obtained from Chinese (21), Philippine (25), and Sikh (24) immigrants in Barcelona. Each respondent (ego) has answered four types of questions:
- questions about ego, including country of origin, years of residence, age, gender, religion, reasons for migrating, ...
- alters a list of 30 persons known to ego; the alters are the nodes in the personal network
- questions about alters including country of origin, country of residence, age, type of relation to ego, ...
- alter-alter ties (undirected) pairs of alters that know each other (according to the respondent)
Alter names have been replaced by numerical ids (0,1,...,29) and ego names by numerical ids precedeed by the terms chinese, filipinos, or sikhs, depending on the community.
The data can be downloaded in the file Signos_public_data.zip (right-click and choose save link as). To follow the steps outlined in this tutorial you should download and extract (unzip) this file on your computer.
signos_public_data contains two study definition files
.ego files define the questionnaire, i.e., the questions and (if applies) a list of potential answers. In the
interviews-directory there are three subfolders
sikhs containing the interview files (
*.int) for the three communities. Each
.int-file contains the anwers of one respondent and, thus, defines a personal network. Most interviews have been conducted with the
signos.ego questionnaire; few with the
signos_p_piloto.ego. (This distinction is only relevant if you open the interviews with the EgoNet software; not for the EgoNet2GraphML converter.)
The EgoNet2GraphML software
EgoNet2GraphML is a software to convert EgoNet interviews into GraphML files and to cluster, aggregate, and visualize collections of personal networks using the methodology proposed in: Ulrik Brandes, Juergen Lerner, Miranda J. Lubbers, Chris McCarty, and Jose Luis Molina "Visual Statistics for Collections of Clustered Graphs". Proc. IEEE Pacific Visualization Symp. (PacificVis'08), 2008 (link to pdf).
EgoNet2GraphML can open one or more interview files that have been filled out according to the same questionnaire (
.ego-file). Then it can export these personal networks to GraphML files either on the individual level or after an attribute-based clustering.
- When exporting networks on the individual level, the resulting network has a node for each alter and it stores the ego attributes as network level attributes and the alter attributes as node attributes. Typically, the respondent has to evaluate the relation between every undirected pair of alters; in this case the resulting network is complete and the alter-alter responses are encoded in link attributes.
- When exporting networks on the class level the user first have to specify how the alters should be partitioned based on alter attributes. Then the class-level networks are exported to GraphML files containing a node for each class and a link for each undirected pair of classes. Summary statistics for the classes (such as number of actors in the class or number of links connecting actors in the same class) are stored as node attributes, statistics for pairs of classes are stored as link attributes.