Urban planning community | #theplannerlife

# Thread: Connectivity index

1. ## Connectivity index

Does anyone have experience using a "connectivity index" to evaluate the degree of street interconnectedness? (i.e., connectivity index = # links / # nodes) How do you handle freeways and associated ramps?

2. Here's a book for you: The Mathematics of Georgraphy by Cambridge (or is it Oxford?) university press. There are several connectivity indexes. Alpha, Beta, Gamma etc. shown in this book. I'd have you mark your ramps as nodes(this is, of course, easier, when ramps are in a clover-leaf shape. Nodes are marked whenever a choice is available. Interchanges between Freeways , ramps onto arterials etc. Connectivity indexes are only really useful when comparing various systems. E.g. "LA's freeways have a higher connectivity (0.146) than San Diego's (0.132)" Also a warning--they will always look low because perfect connectivity means that every node connects with every other node via just one link.

Measurements of Accessibility are often better, in my opinion:

Count and make a table (manually if you have to, but I think there might be some programs out there that could do this) which links are used, how many times to get from every node to every other in the most direct way. Lable your nodes A,B,C,D and your links 1,2,3 for example. (A TO B uses link 1, A to C uses 2 and 1 etc.) This way you can see where traffic is likely to bottleneck, for example, if one link is much more accessible than others. If there is more than one equally short path link-wise between nodes you might want to mark them. A link marked as having high accessibility, but demonstrates significantly less use than other comparable ones in reality, might suggest that there are problems with the link that makes it much less preferrable. Perhaps it is badly potholed or the scenery is too ugly. And resources would be well spent improving these stretches with high-potential.

Finally, remember that any index or measurement is a model. You could mark every single "choice" as a node or only major interchanges if you are only interested in the "grand schemes." Also, the occasional longer than average links can distort your model especially when measuring accessibility. Hope I've helped some with my basic knowledge. Definetly check out the book.

3. Thanks for your help! I will definitely get hold of the book. It sounds like the connectivity index will be most useful for the task immediately at hand because I am comparing urban and suburban systems rather than looking for specific areas of poor accessibility.

4. Glad I could be of help. But I fear that the book might be out of print. Here is a website I found that shows how to calculate the connectivity index. A town is requiring developers to meet a minimum connectivity in their subdivisions:

http://www.townofcary.org/agenda/cou...cm11499e1b.htm