I came into this paper not knowing too much about network analysis, but having some general notion of it through its ubiquity in geographic and neuroscience literature (network distance, social networks, neural networks). I thought the paper did a good job of outlining the fundamentals of the field before progressing into geographic specificities and future challenges. I learned that the most The basis of describing networks is in their topological qualities; namely connectivity, adjacency, and incidence, which is what makes it applicable to such a diverse range of phenomena.
Curtis states that “In some cases these network structures can be classified into idealized network types (e.g., tree networks, hub-and-spoke networks, Manhattan networks.” Are idealized network types simplifications of the input data which are performed to fit a certain standardized model?
On page 104, Curtis mentions that “The choice of network data structure chosen can profoundly impact the analysis performed”, just like scale can influence whether or not clusters are observed at a certain resolution and the choice of some variables over others can influence classification algorithms in SDM. Again, we see that the products of any geographic modeling/ network analysis are not objective, but dependent on subjective choice which requires justification.
I assume that the “rapid rendering” discussed in reference to non-topographic data structures is because of function of quicker run time.Why are the data in non-topographic networks processed more quickly than in topographic ones? Is it because without having to assess relationships between points, each point only has to be accounted for once without regard for its connectivity with other points?
It was interesting to note that one of the biggest challenges or paths forward for geographical network analysis was in applying existing algorithms from different fields to geographic data. Usually the challenges are in adapting current methods for new data types or resolving some gaps in domain knowledge, but this is a different kind of challenge probably born out of the substantial developments made in network analysis in different fields.
-FutureSpock