I was initially impressed with Mappoint 2006, however I noticed that when mapping certain postal sectors that the wrong geocode analysis is performed on the postcode, for example when I map GL postal sectors onto the map of UK a few pushpins appear in DN postal sectors. I imported 50 GL postal sectors and 3 geocoded to DN postal sectors.

Also, when you zoom into a village or town, you may notice that a place name is in view, however when you zoom out the place name is geographically many miles/km away. This means that if someone reading the map doesn't know the area they will think that the place name is close to the village or town, which it is not.

These issues relate to Mappoint 2006 Europe version 13.00.18.1200 and if anyone has found a solution to these issues please let me know, many thanks.

Hi,

Can you explain what a GL or a DN postal sector is ?

The GL that I refer to is the Area, so I refer to these as GL postal sectors. A postcode of GL3 4AB (actual format GL 3 4AB) is broken down as: the GL is the Area, the 3 is the District, the 4 is the Sector and AB is Postcode. There are another 2 characters that are not displayed that identify a specific address, because a postcode identifies approx 20 addresses.

The full list is below, when copied and pasted into Mappoint 2006 Europe I noticed some DN postal sectors appear which is for Doncaster, which is about 110 miles away from Gloucester:

Postal Sector
GL1 1
GL20 6
GL20 5
GL19 4
GL7 7
GL4 8
GL4 6
GL4 5
GL4 4
GL20 8
GL4 0
GL50 1
GL3 3
GL3 2
GL3 1
GL2 9
GL2 5
GL2 0
GL1 5
GL1 4
GL1 3
GL1 2
GL4 3
GL52 2
GL54 4
GL53 8
GL53 7
GL53 0
GL52 9
GL52 8
GL52 7
GL52 6
GL20 7
GL52 3
GL54 5
GL51 9
GL51 8
GL51 7
GL51 6
GL51 4
GL51 3
GL51 0
GL50 4
GL50 3
GL50 2
GL52 5

I’ve tried a few things with this. If you import data via the Data Import Wizard, using the postcode sector as the geographical reference, I think there are around 450 sectors (5%) that result in mistakes ie. like the above example GL1 1 is assigned to DN3 1. Interestingly if you use Data Mapping Wizard to enter MapPoint data eg. population by postcode sector, then draw a 25km radius around Gloucester (the GL area) and export the data to Excel then everything looks OK – GL1 1, GL1 2 and GL1 3 are all included. However if you then read that data back into MapPoint and plot it again the mistakes re-appear.

It seems to be a UK issues. I couldn’t see any problems with the Code Postal in France for example.

Suggestions welcomed as this is the fundamental geography for a lot of analysis.

Thanks for this, at least we can get around the problems in the short term.

Regards,

Stu.

After a few more tests, and using red pushpins so that the postal sectors are clearly mapped I still see the DN postal sectors being incorrectly mapped. I'll run a few more tests and see if there is a workaround.

Regards,

Stu.

Hi,

I bought Mappoint 2006 and encountered the same problem when importing data for territories; 5-10% of postal sectors are mistmatched (I noticed this first when TQ3 appeared in Peterborough.) I got some Bath postcodes near London and I think GL postcodes were mismapped as you have described.

I have found this happens with imported datasets only: if you use the getLocation() function of the API it seems to work more accurately. I searched a long time for a work around but this is all I could find (and using the API this way is a lot slower than importing a dataset from a file.)

I called Microsoft tech support and they said that this is a known issue. They agreed to refund the product so I sent it back to Microsoft and they have been thoroughly useless since. BTW: that was several weeks ago and at the moment: I have no product; I have no refund.

Cheers,
Sam

I found other problems with the Mappoint API (that's Application Programming Interface for those of you smart enough to not get into software engineering ...

It seems that the array of fields stuff for the import methods don't work at all. This is the array that you are supposed to use to tell mappoint how to interpret the columns in the data file you are importing; e.g. to tell it which column holds the territory names, whether the data is postcode, postal area, postal sector etc...

As far as I can see (and I tried every which way of passing an array of variables to this flaming dll) this parameter is not processed at all; it seems to be completely ignored. This suggests that the import functions are only partly implemented in this release. As a software engineer, this rings major alarm bells with me... this stuff doesn't seem to do what it says on the tin.

£250 piece of software to map geo data. Get's significant percentage of mappings wrong? API apparently not fully implemented?

Am I just getting fussy in my old age or does anyone else think there's something going wrong here?

Cheers,
Sam

Hi,

I have spoken to a Microsoft Technician who has stated that there have been many reports of this issue and that a patch is not planned at this time but that these issues maybe resolved within a new version of the software.

The map side of things is fine, I like this interface and what it can do alot. However, the data company used to generate the backend data that geocode imported datasets in the UK was changed last year. Basically this means we are unable to 'trust' the import wizard at all and the situation gets worse:- if we want to map a large area, say the county of Essex, then the mapped information could be disguising errors which we would never know about unless we check the data carefully and in this case we may as well create the territory manually to begin with.

The Microsoft Technician did state that Microsoft would refund us for our purchase. I have raised this with our Corporate software supplier who will take this further with Microsoft.

Hopefully a patch will be released soon and we will not have to wait for a new version of Mappoint.

Regards,

Stu.

In addition to the problems with postcode sectors, I have just created a number of territories from my own data, using postcode districts. (UK Postal Geography). All the districts were located in SE England so was surprised to find I had bits of territory everywhere across the country. From further investigation it appears that MP Europe 2006 allocates around 300 (11%) postcode districts to the wrong location. This seems very unhappy for anyone who wants to use the package as a serious business analysis tool.