Eircode and Google Maps Working

MBSnr

Alun wrote: »

Are you trying to say that's a good thing, because it doesn't seem that way to me. If I'm in Dublin and want to navigate to your house, why should I first have to focus my map on Mayo, assuming I even knew you lived in Mayo, to be able to look up the eircode, or am I missing something?

No I'm not saying it's good! Previously if focused in Galway it resolved to H65. I was just saying it wasn't consistent and jumps to the 'local' 4 digit equivalent which is incorrect.

sondagefaux

It seems to be a spacing issue - when I leave out the space between the routing key and the last four characters I get consistently correct results, even when requesting directions.

If I put a space between the routing key and the last four characters, it can be hit and miss, especially when requesting directions.

sondagefaux

Alun wrote: »

OK, for one I looked up it said it couldn't find it, but gave 3 other "suggestions" all called "Ireland" All 3 had the same last 4 digits, but different first 3, i.e. not A98, but A94, A96 and D14.

On further inspection there seems to be a problem with lookups where either the same eircode is used for multiple addresses at the same location, or one location has multiple eircodes such as Bray Civic Offices.

For example A98 X6N3 fails in Google Maps but returns the following in the eircode finder

INTREO CENTRE BRAY
CIVIC OFFICES
MAIN STREET
BRAY
CO. WICKLOW

But if you enter Civic Offices, Main Street, Bray into eircode finder you end up with a whole raft of possibilities, all with different eircodes but the same location on the map.

Conversely, A98 YX67 also fails in a Google search, but when entered into the eircode finder returns 3 possible addresses, i.e.

The Eircode entered is associated with the following address and business. Please choose an option below:

UNIT 4, INDUSTRIAL YARNS COMPLEX, DUBLIN ROAD, BRAY, CO. DUBLIN
EVEREST CYCLES, UNIT 4, INDUSTRIAL YARNS COMPLEX, DUBLIN ROAD, BRAY, CO. DUBLIN
PIZZARO DEVELOPMENTS LTD., UNIT 4, INDUSTRIAL YARNS COMPLEX, DUBLIN ROAD, BRAY, CO. DUBLIN

again all pointing to the same physical location.

I put in A98 YX67 into Google Maps (with space between routing key and last four characters). It asks me 'Do you mean:' with two clickable options, both labelled 'Ireland'. The top option brings me to V92 YX67 (Lixnaw, Co. Kerry), the bottom option brings me to V93 YX67 (Firies, Co. Kerry).

Putting in A98YX67 (no space) gets no result. There are lots of other units, including a Lidl supermarket, in the Industrial Yarns Complex. Each unit (e.g. Unit 5, Industrial Yarns Complex) has its own Eircode.

Complex clusters of multiple Eircodes at the same physical location like that (other examples would be some shopping centres, e.g. Wilton Shopping Centre in Cork, some office buildings, and apartment buildings) obviously present difficulties for Google Maps.

They're going to have to work out a system whereby multiple Eircodes can resolve to the same co-ordinates, eventually showing the different unit/apartment numbers when they've sorted out the labelling of Eircode locations.

sondagefaux

Alun wrote: »

OK, for one I looked up it said it couldn't find it, but gave 3 other "suggestions" all called "Ireland" All 3 had the same last 4 digits, but different first 3, i.e. not A98, but A94, A96 and D14.

On further inspection there seems to be a problem with lookups where either the same eircode is used for multiple addresses at the same location, or one location has multiple eircodes such as Bray Civic Offices.

For example A98 X6N3 fails in Google Maps but returns the following in the eircode finder

INTREO CENTRE BRAY
CIVIC OFFICES
MAIN STREET
BRAY
CO. WICKLOW

But if you enter Civic Offices, Main Street, Bray into eircode finder you end up with a whole raft of possibilities, all with different eircodes but the same location on the map.

Conversely, A98 YX67 also fails in a Google search, but when entered into the eircode finder returns 3 possible addresses, i.e.

The Eircode entered is associated with the following address and business. Please choose an option below:

UNIT 4, INDUSTRIAL YARNS COMPLEX, DUBLIN ROAD, BRAY, CO. DUBLIN
EVEREST CYCLES, UNIT 4, INDUSTRIAL YARNS COMPLEX, DUBLIN ROAD, BRAY, CO. DUBLIN
PIZZARO DEVELOPMENTS LTD., UNIT 4, INDUSTRIAL YARNS COMPLEX, DUBLIN ROAD, BRAY, CO. DUBLIN

again all pointing to the same physical location.

I've got Google Maps to add Industrial Yarns Complex as a location, although I used its postal address, which is Dublin Road, Bray, Co. Wicklow, even though it's geographically in Co. Dublin (or Dun Laoghaire-Rathdown if you want to be fussy).

Hopefully, they'll now start linking the different Eircodes for the units within the complex to its location, especially now that I've sent them details of all the different Eircodes for the different units within the complex.

Bray Head

I am abroad right now. I logged out of gmail and cleared cookies.

Inserting purely an eircode gets some correct pins and some null responses.

Once I'm logged into gmail it 'remembers' that I have a link to Ireland and all these eircodes return the correct location.

sondagefaux

Bray Head wrote: »

I am abroad right now. I logged out of gmail and cleared cookies.

Inserting purely an eircode gets some correct pins and some null responses.

Once I'm logged into gmail it 'remembers' that I have a link to Ireland and all these eircodes return the correct location.

I've done the same and I get very few null responses - the main ones seem to be from locations where there are clusters of Eircodes in the same building/set of buildings - apartment blocks, shopping centres, office blocks, industrial units etc.

I think Google Maps may be struggling to cope with the concept of the same physical location having multiple post codes.

clewbays

sondagefaux wrote: »

I think Google Maps may be struggling to cope with the concept of the same physical location having multiple post codes.

The idea is to display the location not the text address so they just need to take the first occurrence in the ECAD file of the seven character Eircode irrespective of whether or not there are other occurrences and plot the associated X/Y coordinates.

sondagefaux

clewbays wrote: »

sondagefaux wrote: »

I think Google Maps may be struggling to cope with the concept of the same physical location having multiple post codes.

The idea is to display the location not the text address so they just need to take the first occurrence in the ECAD file of the seven character Eircode irrespective of whether or not there are other occurrences and plot the associated X/Y coordinates.

If multiple Eircodes have more or less the same co-ordinates (e.g. each apartment in a block with its own separate Eircode), then Google Maps will have to find some method of making sure that inputting all of these Eircodes into Google Maps produces the correct result.

At present, it seems that almost none of the locations, such as apartment blocks, shopping centres, office blocks etc, which use multiple Eircodes have any of the Eircodes associated with them producing results on Google Maps.

There must be several thousand such locations across Ireland.

If people can't see results for Eircodes in those locations, they might (incorrectly) conclude that they can't use Eircodes with Google Maps at all.

Google Maps needs to ensure that all Eircodes give the correct results, all the time, or people just won't bother.

I presume they've paid to use Eircodes with Google Maps so they'll make the effort to get it right.

alentejo

Any limit on the number of lockups using Google maps.

ukoda

alentejo wrote: »

Any limit on the number of lockups using Google maps.

None

byrnefm

sondagefaux wrote: »

At present, it seems that almost none of the locations, such as apartment blocks, shopping centres, office blocks etc, which use multiple Eircodes have any of the Eircodes associated with them producing results on Google Maps.
If people can't see results for Eircodes in those locations, they might (incorrectly) conclude that they can't use Eircodes with Google Maps at all.

Yes, I've noticed the same - the Eircode for my cousin's apartment in south Dublin isn't returning any result on Google Maps... Still, it's early days yet, I guess, until an official announcement is made.

NinetyForNone

So assuming its included in the Google maps API, does it stop the need to purchase address data from an other Eircode provider?

sondagefaux

alentejo wrote: »

Any limit on the number of lockups using Google maps.

No. None with the AutoAddress app either...

ukoda

NinetyForNone wrote: »

So assuming its included in the Google maps API, does it stop the need to purchase address data from an other Eircode provider?

It depends on what you want to do,lots of companies will want to purchase the data sets for their own purposes

clewbays

sondagefaux wrote: »

If multiple Eircodes have more or less the same co-ordinates (e.g. each apartment in a block with its own separate Eircode), then Google Maps will have to find some method of making sure that inputting all of these Eircodes into Google Maps produces the correct result.

You enter an Eircode into GoogleMaps. The X/Y associated with that Eircode is also associated with other Eircodes. GoogleMaps ignores that the X/Y is not unique within the ECAD file and just displays the X/Y associated with the Eircode that you inputted on a map. Another way to say this is just display the X/Y associated with the Eircode and don't check whether other Eircodes have the same X/Y. Should be a very simple fix.

BailMeOut

From my testing If a business listing on Google already had the Eircode on the address then pin pointer will exactly match when you do the lookup. i.e. Google will point to where the business put the marker when then set it up.

plodder

clewbays wrote: »

You enter an Eircode into GoogleMaps. The X/Y associated with that Eircode is also associated with other Eircodes. GoogleMaps ignores that the X/Y is not unique within the ECAD file and just displays the X/Y associated with the Eircode that you inputted on a map. Another way to say this is just display the X/Y associated with the Eircode and don't check whether other Eircodes have the same X/Y. Should be a very simple fix.

If lat/long coordinates are used as the unique key for places in the underlying database, then it could be hard to fix. They might have to go through the entire ECAD, looking for entries that have the same location, and adjust the co-ordinates slightly to make them all unique. You could do that mechanically though I imagine.

Sam Russell

plodder wrote: »

If lat/long coordinates are used as the unique key for places in the underlying database, then it could be hard to fix. They might have to go through the entire ECAD, looking for entries that have the same location, and adjust the co-ordinates slightly to make them all unique. You could do that mechanically though I imagine.

The Eircode is unique, the lat/long values need not be. So it is not necessary to continue the search once a match on the Eircode is found - it matters not that other Eircodes use the same lat/long.

plodder

Sam Russell wrote: »

The Eircode is unique, the lat/long values need not be. So it is not necessary to continue the search once a match on the Eircode is found - it matters not that other Eircodes use the same lat/long.

That's how ECAD is structured, but not likely google's places database. There is no precedent for a postcode to be a unique key like this. So, to fit in with google's model, they would have to adjust things. Either that or google adjusts their schema to take account of unique postcodes, but I don't see that happening. [edit] actually looking at their web apis in this area, they seem to have the notion of a place ID which is a long string of letters. So, it doesn't look like lat/long needs to be unique. Not sure what the problem is then.

ukoda

plodder wrote: »

That's how ECAD is structured, but not likely google's places database. There is no precedent for a postcode to be a unique key like this. So, to fit in with google's model, they would have to adjust things. Either that or google adjusts their schema to take account of unique postcodes, but I don't see that happening. [edit] actually looking at their web apis in this area, they seem to have the notion of a place ID which is a long string of letters. So, it doesn't look like lat/long needs to be unique. Not sure what the problem is then.

I wouldn't think it's actually a problem, they are probably just working their way through thier tasks.

plodder

ukoda wrote: »

I wouldn't think it's actually a problem, they are probably just working their way through thier tasks.

Whether you want to call it a problem or not, this aspect of it isn't working.

ukoda

plodder wrote: »

Whether you want to call it a problem or not, this aspect of it isn't working.

No one said it is working

Sam Russell

plodder wrote: »

That's how ECAD is structured, but not likely google's places database. There is no precedent for a postcode to be a unique key like this. So, to fit in with google's model, they would have to adjust things. Either that or google adjusts their schema to take account of unique postcodes, but I don't see that happening. [edit] actually looking at their web apis in this area, they seem to have the notion of a place ID which is a long string of letters. So, it doesn't look like lat/long needs to be unique. Not sure what the problem is then.

From what is being reported here, there is a problem with the space. If it is omitted then it works. There appears to be a second problem with the the random bit being linked to the wrong routing code. A third problem appears to be associated with multiple Eircodes at one location.

Not a difficult list to debug I would have thought.

plodder

Sam Russell wrote: »

From what is being reported here, there is a problem with the space. If it is omitted then it works. There appears to be a second problem with the the random bit being linked to the wrong routing code. A third problem appears to be associated with multiple Eircodes at one location.

Not a difficult list to debug I would have thought.

You're not allowed to call it a problem, don't you know? Or maybe, it's just me.

BailMeOut

plodder wrote: »

You're not allowed to call it a problem, don't you know? Or maybe, it's just me.

in software we call these 'bugs'! It's not even an alpha release and you all doing a fine job helping Google identifying issues that they will fix.

Will be interesting how long they will work on this project before officially announcing it.

ukoda

plodder wrote: »

You're not allowed to call it a problem, don't you know? Or maybe, it's just me.

A problem would be more like insisting that Google couldn't possibly get a licence that wasnt available to other customers. Which turned out to be a non existent problem

plodder

ukoda wrote: »

A problem would be more like insisting that Google couldn't possibly get a licence that wasnt available to other customers. Which turned out to be a non existent problem

or what you wrote in post #23 on this thread.

There's defiantly an issue with the spaces. I just had the same problem, the eircode I entered changed the routing key to the unique indentifier that's closer to me. But removing the space gets me the right location

ukoda

plodder wrote: »

or what you wrote in post #23 on this thread.

It's a problem for us as it doesn't work. But for Google, I doubt they are sitting around scratching their heads, my assumption is they just haven't reached that part of implemation yet. Hopefully you can see the difference and not try start an arguement over it.

BailMeOut

ukoda wrote: »

It's a problem for us as it doesn't work. But for Google, I doubt they are sitting around scratching their heads, my assumption is they just haven't reached that part of implemation yet. Hopefully you can see the difference and not try start an arguement over it.

and if you want to be part of the solution report any 'problems', (bugs) back to Google. You can do this from the map menu / Send feedback / Other feedback.

Sleeper12

I use sat nav 5 to 8 times per day. Google hasn't let me down yet with eircode