Sign In
Register
Products
Solutions
Resources
Pricing
Partners
Company
Howdy, Stranger!
It looks like you're new here. Sign in or register to get started.
Sign In with RapidMiner
Sign In with RapidMiner
Sign In
Register
Quick Links
Categories
Recent Discussions
Best Of...
Unanswered
Groups
Categories
All Categories
19.8K
Help
442
Knowledge Base
Altair RapidMiner Community
GET HELP. LEARN BEST PRACTICES. NETWORK WITH YOUR PEERS.
Discussion
Cluster Measure for geographical data
Author
Date within
1 day
3 days
1 week
2 weeks
1 month
2 months
6 months
1 year
of
Examples: Monday, today, last week, Mar 26, 3/26/04
Search
💬
0 Comments
🔥
0 Discussions
👤
0 Members
🔌
0 Online
ASK A QUESTION
FIND HELPFUL VIDEOS
Home
›
Help
Cluster Measure for geographical data
magic_al
Member
Posts:
11
Contributor II
July 2013
in
Help
I'd like to cluster some geographical data consisting of longitude and latitude. Is there any cluster measure calculating the distance between such coordinates properly? Is it posible to implement one in an existing clustering algorithm?
0
Answers
JEdward
RapidMiner Certified Analyst, RapidMiner Certified Expert, Member
Posts:
578
Unicorn
September 2013
What you are looking for is the Haversine formula which calculates the distance between two points on a sphere.
(See below for a simple explanation of the formula along with a javascript implementation)
http://www.movable-type.co.uk/scripts/latlong.html
At first glance, you'd need to do (as an example & assuming noone has done so already) one of the following.
1-Change the algorithm for calculating the distance between the cluster points to use this formula in the DBscan operator.
2-Install the R extension & use someone elses implementation
http://pastebin.com/TL81ycx9
3-Calculate the distance between all points using a Groovy script version of Haversine & lots & lots of loops.
http://svn.codehaus.org/grails-plugins/grails-solr/trunk/src/groovy/org/grails/solr/Haversine.groovy
Option 2 looks like the easiest, but that said, this is RapidMiner so it's possible that someone has already done this.
Hope that helps!
0
JEdward
RapidMiner Certified Analyst, RapidMiner Certified Expert, Member
Posts:
578
Unicorn
September 2013
Additionally I've found this that should help also, it just needs converted from PHP which shouldn't take too long.
http://www.appelsiini.net/2008/11/introduction-to-marker-clustering-with-google-maps
Good luck!
0
Sign In
or
Register
to comment.

## Answers

578Unicorn(See below for a simple explanation of the formula along with a javascript implementation)

http://www.movable-type.co.uk/scripts/latlong.html

At first glance, you'd need to do (as an example & assuming noone has done so already) one of the following.

1-Change the algorithm for calculating the distance between the cluster points to use this formula in the DBscan operator.

2-Install the R extension & use someone elses implementation http://pastebin.com/TL81ycx9

3-Calculate the distance between all points using a Groovy script version of Haversine & lots & lots of loops. http://svn.codehaus.org/grails-plugins/grails-solr/trunk/src/groovy/org/grails/solr/Haversine.groovy

Option 2 looks like the easiest, but that said, this is RapidMiner so it's possible that someone has already done this.

Hope that helps!

578Unicornhttp://www.appelsiini.net/2008/11/introduction-to-marker-clustering-with-google-maps

Good luck!