Do different cities should have different thresholds for selection of what is a candidate and what is not?
For instance, we use thresholds for population, entorpy, degree, betweeness, closeness and bus frequency.
Having some city clusters suggesting which levels to adopt should be preferred.
Next step: have some municipality cluster process.
Should we use the value for thresholds or the quartiles?
If we have a betweness typically with this distribution (recalling that values are re-scaled 0-1), when cutting the lower and upper 0.20 is different than cutting the lower and upper 20% (quartile).
The cutting value could also come from the clusters, but we know this is a high-sensitivity variable. Quartiles are preferred?
Example for betweeness cuting of 0.20, the number of candidates can be less or more restrictive.
Two research questions here:
Do different cities should have different thresholds for selection of what is a candidate and what is not? For instance, we use thresholds for population, entorpy, degree, betweeness, closeness and bus frequency. Having some city clusters suggesting which levels to adopt should be preferred. Next step: have some municipality cluster process.
Should we use the value for thresholds or the quartiles? If we have a betweness typically with this distribution (recalling that values are re-scaled 0-1), when cutting the lower and upper 0.20 is different than cutting the lower and upper 20% (quartile).
The cutting value could also come from the clusters, but we know this is a high-sensitivity variable. Quartiles are preferred?
Example for betweeness cuting of
0.20, the number of candidates can be less or more restrictive.