The custom geocoding has all been imported fine. The issue is just that Area Units with no institutions in your source data aren't shown on the map. That's just how filled maps work - they are letting you visualise your data based on the chosen geographic breakdown, so there won't be marks (filled areas) for non-existent data.
To prove that the geocoding is correct I've added another sheet to your workbook where the data source is just a list of all of the Area Units in the country - in fact the file used to import the geocoding. I've zoomed that in on Auckland so you can see the missing areas are there, but it includes the whole country.
You may notice that 4 area units aren't shown (there's a message "4 unknown" at the bottom of the screen), which is because Tableau hasn't been able to distinguish between a couple of pairs that it regards as ambiguous. I'm quite surprised by this - by inspection you can see what the issue is - and it's a case that I would have thought was very common and would clearly also be very easy to resolve automatically. Unfortunately all you can do about those is enter the lat/lon by hand, you don't even get to pick from the options, which would be a more helpful option if it's not going to do it automatically.
Anyway, back to your original problem - which is that you want to show all Area Units even if they don't have any institutions. I can see a couple of options:
You could add some type of dummy entry to your data source, one per missing Area Unit. That would force the map to be filled but you'd have to do some messy stuff to avoid marks showing on the dual axis map and in the legend for the institutions. I vaguely recall doing something like that in the past - but I can't remember how successful that was.
Probably the better option is to take a completely different approach to this. I think what you are really wanting is just a background map broken down to the Ward or Area Unit level with the locations of the institutions shown on top (as opposed to, say, colouring the Area Units by number of institutions, or some other metric, which is more what filled maps are intended for). Tableau 9 introduced some new options for driving background maps from either a WMS server or from the MapBox service. I believe you should be able to load the Ward and Area Unit definitions that way, though I've never tried MapBox and only had a brief dabble with a WMS service years ago, so I don't really know what's involved, I'm afraid.
Auckland Wards RL2.twbx 16.2 MB
Thanks for the clarification.
I had a look at MapBox service as you suggested. It has a function to import shapefiles so this is another option to set up the
required boundaries as use it as a background map in Tableau.
Thanks again for all your help
Would you kindly provide a working link to download TabGeoHack?
I just tried your link and it is already inactive
I tried to add a space in different places of the city data, and find that the space is eliminated during the matching.
This is helpful for effectiveness of the matching.
I guess the auto elimination of space works on geo data side too.
Thus a bad solution might be add something like 1,2.... to help remove the ambiguous.
1 of 1 people found this helpful
The download link in page 1 of this thread still works fine for me - I just tried it. Could you please try it again and let me know if it still doesn't work for you? If not, maybe you could try from another machine.
It's not that access to Dropbox is blocked by your work, is it?
That is all pure Tableau doing the matching. The normalisation that is done can sometimes be unhelpful, though. I have had data where the place names in my data exactly matched the geocoding database and would have given unique matches for all rows, but Tableau removed characters like spaces and resulted in various place names being ambiguous.
I have a rare problem. When I import my csv file into tableau I get this error: Error reading from the file Porirua_Meshblock.csv: Invalid separator '٫' The table "[TableauTemp].[Porirua_Meshblock#csv]" does not exist. Unable to connect to the file "". Check that the file exists and that you have sufficient access privileges for it.
I have this problem for all files even your yml file
Yeah I just realized that dropbox is blocked in my LAN. Sorry for that
That sounds like an issue caused by international settings. I'm afraid these utilities don't handle international settings properly, so you will have to set your machine to US or UK settings (or at least a country for which the comma is the separator in a CSV file). You can just make that change temporarily while you run the utilities. I did write up detailed instructions about that when I first wrote these - I think it's covered in the documentation - otherwise very early in the "Grow Your Own Filled Maps" thread. Failing that, you could look for Robert Mundigl's documentation describing how to use it on Clearly and Simply - I know Robert ran into this because he is German and so had different country settings. He wrote a very clear step-by-step guide about it.
Immanuel: No problem, thanks for confirming.
Thank you so much Richard, I have set my machine to US setting, but the error that I mentioned in my first message was appearing again. I read all the comments about your "Grow Your Own Filled Maps" topic and also Robert Mundigl's documentation. I did this process step by step, but it's not working because of this error. @Richard Leeke
To preface this post, I am extremely green when it comes to Tableau as well as GIS file structure/ manipulation/creation. That said I'm enough of a nerd that I believe I can power-through my knowledge deficiencies.
Really interested in the tabgeohack utility you have created. I'm guessing with 10.0 coming out soon that there will be less of a need for the workaround. That being said, I am working on a project and I want to produce some custom regions and I don't have the luxury of waiting for the next version. I think I understand the basics of the "hack", but I am stalled out on finding shapefiles to modify. I am trying to group US zip codes together to create an area of inclusion and the native shapefiles in tableau are more than adequate. Is it possible to combine the native Tableau shapefiles with tabgeohack or do I need to find an outside source for those postal code shapefiles? If an outside source is required, do you have any suggestions for where to find good data? I have downloaded a file from the Census bureau but I am concerned that the file is not of sufficient detail / some of the postal codes are not included. Any direction would be extremely helpful and much appreciated.
I'm not really sure what to suggest. It still sounds to me as if the issue is something to do with locale, or country settings or something - but I really haven't got that much experience of working with different locales.
The only suggestion I have which might help is that you could try opening the file that reports the error when importing custom geocoding just as a text datasource in Tableau. See if that works or gives a similar error. It might just give more of a clue about what is happening, because when you open a text file you can adjust the locale settings and field separator directly in the data source - so you might be able to experiment to see what settings will work for you.
Here's what I get when I open that file and look at the text file properties:
Let me just make sure I understand exactly what you are trying to do. If I understand correctly, you want to produce custom regions which are composed of groups of US zip codes. Is that right?
If so, how will you determine which zip codes need to be grouped together? Are you seeing that as a manual process - by defining the lists of zips that get grouped into each region? I can imagine doing that by selecting the groups from a map, or by creating a table which groups the zip codes by region code in some way and then running a utility to merge them.
You are right that the ability to group Tableau's built-in shapes in that way is coming in Tableau 10.
In the meantime, I think there are three steps in the process that you need before tabgeohack can help you.
1) Identify a source of zip code boundary data. That isn't something that tabgeohack will do for you, though I do have a vague distant memory of having worked out how to extract Tableau's boundaries into shapefile format once. I really can't remember how I did it or whether I have any tools lying around for that. I'm afraid I have no idea about availability of that data online, though. In some countries that data may be freely available, in others it can be expensive. Google is your friend here - that's where I would have to start.
2) Identify the zipcodes you want to group and then find a tool to do it with. Various free/open source GIS tools have that capability. It is also possible to do it by loading the data into a spatial database and using spatial operations. I've done it in the past with PostGIS (which is the open source Postgres relational database plus GIS extensions) - but I don't have PostGIS running on my current machine and can't remember the details.
3) Save the resulting merged shapes as an ESRI shapefile (or one of the other spatial file formats that tabgeohack supports).
Now you're ready to start importing it with tabgeohack.
I'm not sure when Tableau 10 is due out - but it might well beat you to the punch. :-)