relative GIS newbie here.

I've been working on a personal project to rate roads on how nice a ride they'd be on a motorcycle. I wrote a Python program to work through road data in the TIGER dataset (PostGIS) and apply a small number of metrics:

  1. Road variance (twistyness)
  2. Road elevation variance
  3. Does the road pass by a body of water
  4. Does the road go through a park

The results of applying the metrics have been a good start, but they are far from perfect. So I'm trying to decide on some new metrics to apply.

  1. Check land coverage around road. Riding along rolling hillsides is better than a forest which is better than an industrial park.
  2. Road condition (paved or unpaved)
  3. Number of stop signs, street lights encountered
  4. Utilize viewshed analysis to check if a body of water is actually visible from road
  5. Identify mountain views and use viewshed analysis to see if mountain can be seen from road (think Great Smokey Mountains)
  6. Apply historical traffic data

I need the advice of some GIS professionals. Do these sound feasible, or even make sense? Can you think of any other things I should try?

Most importantly, where can I get data for the valid ideas?

  • they make sense, but some metrics seem too subjective: eg. paved or unpaved road. That depends. If you own a Harley and likes a smooth ride or if you own a trail bike (not sure what the proper name is) that is suited for bumpy rides. Commented May 5, 2011 at 20:30
  • Hi Eric, I've just written a similar python program that evaluates road curvy-ness based on OpenStreetMap data: github.com/adamfranco/curvature/wiki One advantage of OpenStreetMap is that road surface and smoothness can be added to the data set, though for many places this isn't available currently. Get in touch if you are interested in collaborating. Commented Dec 20, 2012 at 16:59

2 Answers 2


From GIS perspective you may analyze any dataset available in almost any way you can think of, so technical feasibility is not an issue here. Espetially that you are familiar with Python (I would recommend QGIS/GRAS soft).

However, it seems like your question is directed to landscape profesional rather than GIS. As from your description it sounds like you want to receive visual perception assessment of the landscape that you are driving through.

I think it is great and complex task to perform such analyze, as if you will base your assessment only on geographical futures, than you will receive only landscape assessment, which is closely related to visual assessment, though these are two totally different animals. What you want to rate (in my opinion) is a personal experience of the driver rather than just presence of physical features.

There are loads of visual, scenic beauty assessment papers in the net, so probably you should start to look for indices in which you can measure attractiveness of the landscape.

You will have to analyze mainly your viewshads extent and complexity (from driver perspective). Commonly considered as attractive are: skylines, landscape complexity, naturalness (which is not exact term), cultural features, extent of view (width, depth). The tricky part is, that not always the most natural is the most attractive, as combination of certain cultural features may be more attractive than semi natural landscape. Usually your visual experience would need to be assessed case by case on site, thats why I think itis a difficult task to realize accurate matrix, though not impassible.

Sorry for being boring and actually with out any valuable conclusion in the end. I'll think if I can think of any practical advices.

  • Thank you. Conclusions are difficult at this stage, but advice that points me in the right direction is very much appreciated right now. You gave me a lot to look into, so thanks! Commented May 9, 2011 at 16:06

This article, which appeared in ESRI's ArcUser Spring 2010 publication, has a great overview of the process they took when trying accomplish what you are trying.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.