Introduction

What is Network Analysis?

Definition

“Networks, or graphs, consist of verticles and edges. An edge typically connects a pair of verticles” (Fortunato & Hric, 2016)

• What can we do with this technique?

e.g. “What are institutions participating in the mining of cryptocurrencies?”

What is Clustering?

• Clustering is a technique used to build communities by grouping relatively closer entities within the networks

  → The aim is to build sub-groups

“Most networks of interest display community structure, i.e., their verticles are organised into groups, called communities, clusters or modules” (Fortunato & Hric, 2016)

• When we map a network, we will not look at causalities
  • That does not mean that we cannot identify which nodes are the most important
  • Nodes are often seen as gatekeeping positions that shape the overall network

Networks of Actors

• two approaches:
  • focus on an R&D area and identify key actors producing STI
  • start from a MNE and see how it is mostly working with to produce STI
• Collaborations are more clearly seen by studying scientific publications rather than patents
• When looking at scientific publications, we build networks based on co-authorships which show actual collaborations (an alternative could be to build a network based on the references used in each paper)

Example 1:

Top 200 HBMS (Health and Biomedical Sciences) research affiliations in the top 30 journals by impact factor (2009-2018)

• limit the number of entities to be plotted, so here only top 200 in a sample that was retrieved from a much larger data set
• We can also plot a third variable (e.g. country)

Example 2: Star Graph with one central actor