Clustering Methods?

[andrew c.]

Hello all. I work here in the US for a state agency using the MapInfo software.

We recieve data for traffic stops on a large scale - the entire state. The number of points can can vary from 50,000 to 100,000 or more.

We would like to determine where the highest frequency of these crashes occur based on a clustering method. (For example, we would like to find the locations where there are 50 or more points within a 25 foot diameter of each other.)

In order to perform this operation, we would have determine the distance between each point and every other point.

We do recieve location data in the form of Latitude/Longitude decimal degrees.

I would like to hear if anyone has a solution for our problem. If there a software plug-in for MapInfo to perform this? Or is there another software available? Or can we just use the Lat/Longs in some kind of statistical program to determine clusters?

Any solutions would be greatly appreciated!

Andrew

[CDR Group]

Hi Andrew

 

Have a look at the Engage 3D or Crime Profiler derivative - both PB products.

Hot spotting and a lot more.

You could start analysing times of day and days of the week also.

http://www.pbinsight.com/products/location-intelligence/applications/mapping-analytical/mapinfo-crime-profiler/

 

Regards

 

John Ievers

CDR Group
www.cdrgroup.co.uk
www.browse-a-map.co.uk
Tel: 01433 621282
Fax: 01433 621292
Specialists in Geographic Information Systems

CDR Group is the trading name of Contract Data Research Ltd.
Registered address Eccles House, Eccles Lane, Hope, Hope Valley, S33 6RW.
Registration No. 1972326
VAT No. GB373 3117 67

--
You received this message because you are subscribed to the
Google Groups "MapInfo-L" group.To post a message to this group, send
email to mapi...@googlegroups.com
To unsubscribe from this group, go to:
http://groups.google.com/group/mapinfo-l/subscribe?hl=en
For more options, information and links to MapInfo resources (searching
archives, feature requests, to visit our Wiki, visit the Welcome page at
http://groups.google.com/group/mapinfo-l?hl=en

[Blair Freebairn]

Hi Andrew,

I am not sure simple surface hot-spotting is the right analytical
framework for your task. Vehicles cannot move freely across the
surface (unless they are a tank!) so the start point for your analyses
should be the road network and your spatial hot-spotting should
respect the topology of the road network (else two parallel but
completely unconnected roads shouldn't be allowed to influence each
other to produce a spurious 2D hotspot. Your problem is essentially a
1D one.

That said if you were looking for a MIP plug in for what you are after
I would suggest Vertical Mapper. However I suspect you will almost
certainly find yourself having to 'roll your own' in a statistically
capable software framework of your choice, SAS, R, and SQL (especially
now has rich spatial in either MSSQL, Oracle or PostGIS) are all
perfectly sensible tools to start building hot-spotting algorithms for
a road network.

Also with crash data you need to think pretty hard about the
denominator for your frequency calcs. Is it crashes/vehicle movement,
crashes/road km, crashes/addresses etc etc


Good luck

On Mar 13, 8:21 am, "CDR Group" <sa...@cdrgroup.co.uk> wrote:
> Hi Andrew
>
> Have a look at the Engage 3D or Crime Profiler derivative - both PB products.
> Hot spotting and a lot more.

> You could start analysing times of day and days of the week also.http://www.pbinsight.com/products/location-intelligence/applications/...
>
> Regards
>
> John Ievers
> CDR Groupwww.cdrgroup.co.ukwww.browse-a-map.co.uk

[Greg]

Andrew,

You might also want to have a look at Crimestat - http://www.icpsr.umich.edu/CrimeStat/.  It's a standalone program but will read MapInfo.dat files and output in MID/MIF and it’s free!  While principally aimed at crime analysis, it has a range of point analysis functions that can equally be applied to other areas, such as yours.   It also has a comprehensive help manual, which explains in detail the functions and options available.

HTH

Greg.

[Glen]

i would start by just plotting the points, and look for groups of
points
can you post the data ?