Declan You should try CrimeStat2. It is much less quirky in conjunction with MapInfo. I rarely had any problems. Dick
Richard Block Department of Sociology Loyola University 6525 N Sheridan Rd. Chicago IL 60626 Telephone 773 508 3454 FAX 773 508 7099 email [EMAIL PROTECTED] >>> Declan Troy <[EMAIL PROTECTED]> 09/09/02 18:38 PM >>> Not too much response to my inquiry, perhaps reflecting an application not too widespread among MapInfo users. No one confirmed that the VM kernel routines work (there was one voice of trepidation from a user planning on using this analysis). CrimeStat 2 was suggested as an alternative. I actually have used Crimestat and can vouch that it does produce much more realistic looking results from kernel analyses than does VM. The price is right (free), the author is very responsive to problems, and is the only source I can think of off hand for dual kernel analysis. Although one can transfer information between MapInfo and CrimeStat, I have found this to be rather quirky. At least Crimestat does allow for transfers to MapInfo (and other formats). The VM, being a MapInfo add-on is much more tightly integrated and easier to use. The Mapinfo/ESRI discussion has an asymptote to this problem. ArcView users have long had access to some nice routines for kernel analyses, probably reflecting the larger user base of academic users. Thanks to those who responded. Declan >Is anyone using Vertical Mapper to do kernel analyses (creating >grids based on point density)? I have two problems when I try. > >First, the contribution of adjacent points seems to be ignored; >i.e., rather than summing densities VM appears to just overlay >densities for each point; i.e., densities of each point erase any >prior points they overlay. Density grids calculated using the >"Square Area" method of point density analysis (which is a specific >type of kernel analysis) look fine. > >Second, one is supposed to be able to attach weighting to points >(event sum column) but again when I try this the results don't look >realistic (points with high values don't correspond to higher >densities in the resultant grid), even for the Square Area analysis. > >If anyone is using VM for kernel analysis and is either getting good >results or can confirm problems with these routines please let me >know. > --------------------------------------------------------------------- List hosting provided by Directions Magazine | www.directionsmag.com | To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] Message number: 2918 --------------------------------------------------------------------- List hosting provided by Directions Magazine | www.directionsmag.com | To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] Message number: 2920
