Thanks for starting this thread on what I believe to be an important topic.
You may be interested in the the work of Willis and her team who have produced maps showing the prevalence of Parkinson's in each of the approximately 3000 counties in the US . There are fascinating differences from region to region which, if we could understand, would take us closer to knowing the etiology of the disease.
I was unable to find something similar for the UK. So I produced my own map (for England) using government data. You can access for each NHS practice the number of prescriptions they have written for each drug in any month. It's not precise, but by associating some drugs, such as levodopa, with Parkinson's you can get a measure of the relative number of PwP in a practice compared with the national average. I calculated these values and plotted them on maps . The next stage is to see what is causing the differences. It could be some real factor or it could be "noise".
The UK government has an open data initiative and are trying to make it easy to access data. See:
If anyone is keen to be involved in studying the epidemiology of Parkinson's, please contact me.
 Neuroepidemiology. 2010 April; 34(3): 143–151.
"Geographic and Ethnic Variation in Parkinson Disease: A Population-Based Study of US Medicare Beneficiaries"
Allison Wright Willis,a,* Bradley A. Evanoff,b Min Lian,b Susan R. Criswell,a and Brad A. Racetteahttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC2865395/