There is basically no limit to the phenomena that can be modeled and analyzed in terms of complex networks - entities and their relationships between which can be represented as the nodes and edges of a graph, and which form a non-trivial pattern. So let's make this a small survey:
- Where in your research do you employ complex networks and network analysis methods?
- What are your data sources? How big are they?
- Which tools do you use for the network analysis process?
- What did you learn from the network analysis?
Where in your research do you employ complex networks and network analysis methods?
In air passenger and air freight transportation
What are your data sources? How big are they?
The US BTS has a data series called T-100. This can provide flows between airports (i.e. link or arc flows). There are also data with the true Origin and Destination of the passenger flows (DB1B). These are quite large and detailed data -- the later is a 10% sample of all passengers. There are some interesting problems interpreting and tying these data together. Other data could come from commercial sources.
Which tools do you use for the network analysis process?
Develop my own algorithms for coding and optimizing hubs, counts and characterizations of their scale free or other measures of robustness, ... and other descriptive aspects of networks
What did you learn from the network analysis?
It is possible to characterize the efficiency or at least to understand the organizational strategy of the companies running these networks, and possibly (although this is work in progress) determine their comparative efficiency in use of aviation fuel.
Article Air freight hubs in the FedEx system: Analysis of fuel use
Christian,
I agree, there are so many systems that can be modeled as networks and then mapped and measured and analyzed. I have been apply network analysis to business issues for > 20 years. See a collection of my client projects here:
http://ognet.com/cases.html
More experimental projects appear in my blog
http://thenetworkthinkers.com
Data sources are everything from employee surveys to spreadsheets, to corporate data bases, to public / open source data bases, to screen-scraping web pages. Projects range from 50 nodes to 5000 nodes usually. Some projects are 50,000 to 500,000 nodes -- which usually get broken up into sub-networks of 100 - 1000 nodes before we start serious analysis. My main focus is human networks in organizations and communities.
I use home built software -- InFlow, which is also a commercial product > 1000 users and SNA. I do most of my data munging using Prolog -- which, as a relational language, is the perfect computer language for networks.
Enjoy reading!
Valdis
Christian,
You might find this SNA study I just presented at the University of Cambridge of interest. We look at the networks that were involved in the Mortgage Meltdown of 2007-2009. Nodes are people, organizations and financial instruments. Links are mostly financial transactions. Data was gathered from public/government sources.
The presentation is here on ResearchGate...
https://www.researchgate.net/publication/266896531_From_Main_Street_to_Wall_Street_Connecting_the_Dots_in_the_Mortgage_Meltdown
Enjoy!
Valdis
With the Ebola outbreak being at in the News all over the world, this article by the CDC talks about applying SNA to outbreaks of contagious diseases -- they call it Transmission Network Analysis.
https://www.researchgate.net/profile/Valdis_Krebs/publication/6775900_Transmission_network_analysis_to_complement_routine_tuberculosis_contact_investigations/links/53d8d2020cf2e38c6331abaf?origin=publication_detail
Article Transmission Network Analysis to Complement Routine Tubercul...
Where in your research do you employ complex networks and network analysis methods?
I am currently looking at online social movements.
What are your data sources? How big are they?
I can end up with fairly large sets of nodes and edges. Hundreds of thousands of actors when unfiltered, although I usually apply some form of sampling or filtering. For Twitter, events can be pretty good network boundary setters. Always depends on the research question of course.
Which tools do you use for the network analysis process?
I have applied ERGM modelling and standard social network descriptive statistics, although my current favourite for large datasets is using correspondence analysis to analyze bipartite networks.
What did you learn from the network analysis?
This is key. With large datasets learning new things can be difficult - at the best of times you mostly see peculiar relationships that can be hard to explain with any level of validity. Often I find myself puttering around with the data to look for a stable pattern that I can explore more deeply, or use as a model to check against other areas of the data. That said, I learned that there is unexpected power and influenced exercised by groups of Twitter users who are simply adept at finding and curating information on their favourite topic. They can be even more influential than celebrities and impassioned activists in certain contexts.
Dear Christian,
Nice idea for a small survey, here are my answers.
Where in your research do you employ complex networks and network analysis methods?
Research in personality psychology.
What are your data sources? How big are they?
In my research, nodes represent personality characteristics (of different levels of abstraction) and edges represent their pairwise relationships. These variables are usually collected by administering questionnaires to a few hundreds of participants up to a thousand. The number of nodes is usually not large (a few tens to one hundred), but we may investigate larger networks in the future.
Which tools do you use for the network analysis process?
Usually R, in particular packages qgraph, igraph, sna, parcor have been particularly useful.
What did you learn from the network analysis?
It's a great theoretical and methodological perspective in my field and we just started adopting it. As an example, we are investigating ways to find which are the most relevant personality characteristics, those that are expected to produce greatest overall changes when manipulated.
Here are a few papers that can give the idea of what we do.
Article Deconstructing the construct: A network perspective on psych...
Article qgraph: Network Visualizations of Relationships in Psychometric Data
Article Generalization of Clustering Coefficients to Signed Correlat...
Article State of the aRt personality research: A tutorial on network...
I would point you to the very dynamic field of complex networks, involving large amounts of nodes and edges, and community detection is particularly active subject. The following paper presented a new way of thinking in community detection or classification in general:
Jiang B. and Ma D. (2015), Defining least community as a homogeneous group in complex networks, Physica A: Statistical Mechanics and its Applications, 428, 154-160.
Jiang B., Duan Y., Lu F., Yang T. and Zhao J. (2014), Topological structure of urban street networks from the perspective of degree correlations, Environment and Planning B: Planning and Design, 41(5), 813-828.
Where in your research do you employ complex networks and network analysis methods?
We use complex networks as underlying structures for different diffusion processes.
What are your data sources? How big are they?
We use mainly artificial networks constructed by some models (e.g., Barabási-Albert model, ranking method for scale-free networks, Watts-Strogatz model), but we also work with real networks representing social structures obtained from facebook or twitter. The networks which we analyse have up to one million nodes.
Which tools do you use for the network analysis process?
We write our own programs in C++.
What did you learn from the network analysis?
So, one of our goals is to identify network characteristics which are crucial for the behaviour of different spreading processes. One of such important factors is the average node degree of a network, but there are also other influential features.
Here is a paper in which you can find more about that:
https://journals.aps.org/pre/abstract/10.1103/PhysRevE.95.012307
Article Pair approximation for the q-voter model with independence o...
I recommend the book "NETWORKS" by Mark Newman, 2018.
NETWORK ANALYSIS IN PSYCHOLOGY Eduardo Fonseca-Pedrero, 2018
Article Análisis de redes en psicología (NETWORK ANALYSIS IN PSYCHOLOGY)
Article Network analysis: A brief overview and tutorial
A History and Applications of Network Analysis - Squarespace
static1.squarespace.com › static › Lecture+4+History+of...
PDFNetwork analysis begins with solution to the “Bridges of. Königsberg” question in 1735. James D. Wilson (USF). History and Applications of Networks. 4 / 20.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs