Instead of explaining it all, I've attached an example. This has the datasource and several examples in a packaged workbook. Note the importance of having all the points in the polygon and the order that those points are supposed to be connected.

Dan,

The Packaged Workbook that Marc attached can be unpackaged to reveal both the data and background images used in the workbook.

To unpackage the file:

1. Right-click the .twbx file in Windows explorer.
2. Select Unpackage.

The resulting folder has a spreadsheet in it called election.xls, which contains an example of how the latitude and longitude values were structured.

However, for convenience, I've attached the MS Excel file below.

One source for Zip code boundary files is the US census, http://www.census.gov/geo/www/cob/zt_metadata.html

These are actually ZCTA (Zip Code Tabulation Area) boundaries, as Zip codes are delivery routes and do not always correspond exactly to a polygonal area. See the Census web site for more discussion of the differences.

The census boundary files may be excessively detailed, causing Tableau to spend a long time rendering them.

Hi guys,

I was just about to ask this question, and realized there is a good dicussion going on.

I am a complete newbie to Tableau, but have mapped data in Arcview. What do I have to fill in polygons here in Tableau - right now, I just want to fil in states.

I know you gave a great example, I just need to be explained the process. Essentially, what data files do I need to bring over from Arcview and how do I attach the data in Tableau?

Many thanks for your insight,

Brock

As an update: I was able to hack through the ascii file to pull the information for Georgia and create zip code level detail. Still looking for a quicker way to convert those census files into Access. It was quite tedious and would not be time effective for all 50 states.

Chris (Cush) - can you explain how you crunched the numbers from ASCII and how you got it into Tableau?

An Access 2000 .mdb file of the zip code data is being uploaded right now (50 MB) to my server. I'll have the link here in about an hour. It has 52 States with 50,080 Polygons with 5,732,964 Points. The compressed file expands to about 400 MB.

Please be aware that this data is at least 9 years old, and that is is not recommended to have Tableau draw all the zip code polygons for all the states at once.

Sometime this week, I plan on making a short video tutorial on how to use this data.

If you have any isuess or comments you can contact me here or via http://twitter.com/joemako (or if you need help reformatting your data for use in Tableau)

Hi Joe,

I know am asking alot considering you already uploaded a database for us to use, but have you had a chance to create a brief tutorial for newbies like me on how to associate data with the polygons?

Thanks again!

Brock

Good idea Brock. I could use a mini write up too on how to use the data. Thanks!

Wow, there is obviously a lot of buzz going on in this area. I'm new to using Tableau and I'm currently doing an R&D engagement with our data. I looked at the elections.xls file in the packaged workbook example provided towards the top of this thread. It has two fields in it that I'm interested in: PolygonId and PointOfOrder. These two fields are obviously critical to creating the Polygon Map example which is very similar to what I'm looking to create. I don't have this data and I'm interested in obtaining it. Currently I've got city, state, zip code, and lat/long coordinates...how do I go about obtaining these last two fields though? Anybody know?

Within the elections data there are 5 states ('NY','MN','VA','MI','TX') that have more than one Polygon Id associated with them. In my smaller subset of data I'm trying to determine which Id to use with each Latitude and Longitude coordinate since my set of coordinates are different from the ones provided in the elections data. It feels like it is a guessing game (i.e. numbers just being close) rather than a science. I wish the technical support at Tableau would provide more insight and direction in this area since the Polygon File Structures require both the coordinates as well as the Polygon Id and Point of Order.

OK, let’s see if I can approach this in another way. Instead of looking at states, let’s take a look at some simple shapes, and relate it to working with states.

Notes on the attached image "layout.png":
- Shapes A and B are single contiguous polygons (Similar to a state like Utah)
- Shape C has two polygons (Similar to a state with an island, like Texas)
- Points are ordered like a connect the dots game, if you change the order, the shape will be different.
- For each polygon, the last point is a repeat of the first point to close the loop (this may not be required, but I think it is a good practice)
- The values on the X and Y line can be used similar to Lat/Long

Notes on the attached Excel file "shapes.xls":
- On the "Shapes" tab, you can see 3 different Shapes (A,B.C), 4 PolygonIDs (1,2,3,4), 19 PointOrders (1-19), and X and Y coordinates for each point. All this data came from the layout.png image. (similar to state polygon/point data in that states are like shapes)
- On the "Data" tab, this would be your source data, Shape A has 1 store located in it, Shape B has 2 stores, and Shape C has 3 stores. I've also included sales figures for each store.
- The "Combined" tab is what you get after merging these two data sets together, (I'll get into how it was made next).

On the "Shapes" tab, 5 rows describe Shape A, 6 describe B, and 8 describe C.

Taking a look at the "Combined" tab, you see 41 rows of data, 5 for Shape A, 12 for B, and 24 for C.

So 5 rows became 5, 6 rows became 12, and 8 rows became 24. The math is (5*1)+(6*2)+(8*3) = 5+12+24 = 41 The 1,2,3 came from the number of stores in the corresponding shape.

Attached is an image "query.png" that shows the two tabs and the query that generated the "Combined" tab.

The query is simply a query with both tables, joined on Shape that include the * from Shapes and the fields you want from your data.

When we have Tableau connect to the Combined data source, we are able to see the result, attached as "Shapes.twbx".

I made a couple of tabs that provide different views of the three tabs.

If any of this does not make sense, please let me know what part(s) I was not clear on.

I have also attached a "States.xls" that came from the elections.xls file that has the same structure as the "Shapes" tab (State abbr. instead of Shape, and Lat/Long instead of X/Y)

Best Regards,
Joe Mako

Hi All,

Tableau geo data is very much geared towards US. Does anyone know where I can get country polygones, or even internation cities such as London lat/lon coordinates.

Has Tableau decided to address this yet? It is frustrating to have such a powerful program be limited by a seemingly fixable problem? The anon user from the 25th is right on the money.

BUMP.

I'm trying to do the same thing - figure out / extract the point orders from the census data.

joemako,

Excellent post explaining how point orders work.