Social Network Analysis for Behavioural Scientists using R (SNAR01)
2 July 2018 - 5 July 2018£220.00 - £530.00
This workshop will provide students with the opportunity to learn how to use social network analysis to analyze social relational datasets such as human friendship networks or animal social networks. Attendees will learn how to use R and several R packages including igraph, sna, network, asnipe, timeordered, tsna to create network graphs, calculate descriptive network metrics, use randomization and random models to evaluate the significance of these metrics, determine graph structural properties including community structures, use QAP and MRQAP methods to assess how network characteristics relate to other individual and relational attributes, and measure change over time in dynamic networks. Attendees will also learn how to produce high quality network visualizations using R.
This workshop is ideal for any behavioral scientists including psychologists, ethologists, behavioral ecologists, behavioral neuroscientists, economists, who are interested in learning how to collect data for social network analysis or already have data that they wish to analyze using these methods.
Venue – PS statistics head office, 53 Morrison Street, Glasgow, G5 8LB – Google map
Availability – 30 places
Duration – 4 days
Contact hours – Approx. 28 hours
ECT’s – Equal to 3 ECT’s
Language – English
We offer COURSE ONLY and ACCOMMODATION PACKAGES;
• COURSE ONLY – Includes lunch and refreshments.
• ACCOMMODATION PACKAGE (to be purchased in addition to the course only option) – Includes breakfast, lunch, welcome dinner Monday evening, farewell dinner Wednesday evening, refreshments and accommodation. Self-catering facilities are available in the accommodation. Accommodation is approximately a 6-minute walk from the PS statistics head office. Accommodation is multiple occupancy (max 3-4 people) single sex en-suite rooms. Arrival Sunday 2nd July (after 5pm) and departure Thursday 5th July (accommodation must be vacated by 9am). An additional nights accommodation can be purchased, departure 9am Friday morning email for details.
Other payment options are available please email firstname.lastname@example.org
Cancellation policy: Cancellations are accepted up to 28 days before the course start date subject to a 25% cancellation fee. Cancellations later than this may be considered, contact email@example.com Failure to attend will result in the full cost of the course being charged. In the unfortunate event that PS statistics must cancel this course due to unforeseen circumstances a full refund for the course will be credited. However PS statistics cannot be held responsible for any travel fees, accommodation or other expenses incurred to you as a result of the cancellation.
Introductory lectures on the concepts and refreshers on R usage. Intermediate-level lectures interspersed with hands-on mini practicals and longer projects. Round-table discussions about the analysis requirements of attendees (option for them to bring their own data). Data sets for computer practicals will be provided by the instructors, but participants are welcome to bring their own data.
Assumed quantitative knowledge
A good understanding of statistical concepts, statistical significance and hypothesis testing.
Assumed computer background
R experience is desirable but not essential. Attendees ideally should be able to import/export data, understand basic R syntax and write simple functions and loops.
Equipment and software requirements
A laptop/personal computer with a working version of R or RStudio. R and RStudio are supported by both PC and MAC and can be downloaded for free by following these links.
It is essential that you come with all necessary software and packages already installed (you will be sent a list of packages prior to the course) as internet access may not always be available.
UNSURE ABOUT SUITABLILITY THEN PLEASE ASK firstname.lastname@example.org
Meet at 43 Cook Street, Glasgow G5 8JN at approx. 17:00 onwards
Monday 2nd – Classes from 09:00 to 17:00
Module 1: Introduction to Social Networks Theory. This will cover central themes of social network analysis: the major data formats, structures and collection methods, the different types of graphs and networks; how to generate and visualize social networks and generate basic descriptive statistics, and what hypotheses and questions can be addressed using social network analysis. We will also discuss data types and questions of interest to attendees.
Module 2: R refresher and R packages. This module will provide a quick overview of the major packages used for social network analysis in R including ‘igraph’, ‘sna’, ‘network’. We shall learn how to convert raw data formats to network objects in R; how to interface with R network objects and how to create simple network visualizations.
Module 3: Intro to Visualizing Networks. We shall cover how to generate and beautify networks using the ‘igraph’ R package covering issues such as layout decisions, coloring and sizing of nodes and edges by network attributes, metrics or community. We shall extend this to cover how to create dynamic three-dimentional network plots using the R package ‘threejs’. We shall also discuss how to use the ‘ggplot’ based ‘ggraph’ R package which has many customizable features for plotting networks.
Tuesday 3rd – Classes from 09:00 to 17:00
Module 4: Identifying important nodes and edges. Learn how to evaluate key indicators of each node’s significance to the network including, degree centrality, eigenvector centrality, power centrality, and betweenness. Describe how to calculate for directed vs. undirected and weighted vs unweighted networks. Learn how to assess simple relationships between nodes including geodesic distances, identifying neighbors, determining shortest and longest paths.
Module 5: Introduction to Network Randomization and Random Models. It is critical in network analysis to evaluate how likely it is to observe a given network metric for our network of interest. This module will introduce how to use null models (randomizations/permutations or random graphs) to test whether networks have characteristics that are especially surprising after accounting for non-independence. We also will learn how to generate confidence intervals for network metrics and carry out significance testing. We shall learn how to use the ‘igraph’ package for random graph generation.
Module 6: Network Graph Characteristics. We shall cover concepts such as dyad and triad censuses, transitivity, assortativity, homophily, reciprocity, clustering and density. We shall discuss their significance and importance for answering relevant questions to the patterns of social associations and behavior in networks.
Wednesday 4th – Classes from 09:00 to 17:00
Extending Network Analysis.
Module 7: Community Detection. Overview of what communities (modules) mean for animal and human social networks – that a high proportion of nodes or edge weights cluster within a sub-group of nodes/edges rather than between sub-groups. We shall review the numerous community detection methods and describe the implementation of major ones in R. How to generate robustness in evaluation of community detection. How to to determine the degree of community structure in a network using the index of modularity (Q) and bootstrapping techniques such as community assortativity (rcom). Hierarchical clustering for analysis of hierarchically organized social societies.
Module 8: Randomizations and Random models II. This module will further explore how to determine the appropriate choice of null models for behavioral data. This is not always a trivial exercise for behavioral datasets. We will use the ‘asnipe’ R package for network permutation and ‘igraph’ R package for null model generation. We shall also cover options for dealing with missing data, low sampling rates and pseudo-replication, options for data imputation, and how to account for temporal structure in data randomizations.
Thursday 5th – Classes from 09:00 to 17:00
Module 9: Quadratic Assignment Procedure (QAP) Regression. Using QAP regression to control for non-independence of data when comparing network position or strength between networks or comparing individual/dyadic network metrics with other individual/dyadic attributes. Extending analyses to using multiple regression quadratic assignment procedure (MRQAP). How to perform in base-R and using the ‘asnipe’ R package. How to generate effect sizes when using QAP and MRQAP.
Module 10: Visualizing Networks II. This module will tackle advanced options for network plotting, including how to export ‘igraph’ R objects to Gephi for generating even more beautiful customized plots, how to create interactive web based network visualizations using R packages such as ‘threejs’, ‘visNetwork’ and ‘networkD3’, and how to plot or animate dynamic social networks.
Module 11: Dynamic Networks. Key questions that are often neglected are how do network structures remain stable or change over time and can we infer how meaningful this stability and instability is? This module will discuss various methods for analysis of change for time-ordered and time-aggregated networks. We will use R packages for analysis of dynamic networks including ‘timeordered’, ‘networkDynamic’ and ‘tsna’.