Maps - Relationships with 2 files for UK postal code

[Anonymous]

Hi All

I have been fighting with maps for some time trying to get what I thought was a simple heat map to work based on UK postcodes.

After a bit of reading, I found a user who used the Office for National Stats and downloaded all post codes with the long\lat for each, which is a huge file, but now i want to match the post code from my data file with the national stats post cpde and need to find the best way to do this?!

Can you advise how best to achieve this please....then how best to show a count of people in a specific post code on to a map.....

Thank you

[davegwynedd · Accepted solution]

I've been using the various post code files from National Statistics and there are usually a number of different post code columns in the files. I'm assuming you've transformed the data in some way?

I can see 3 post codes in the NSPL_AUG_2019_UK file - headers: pcd, pcd2 and pcds. We had an issue with using the first two on our datasets but matched pretty well on pcds. Should be the same on the Nov release.

This assumes your post code data is clean - and that it's all in the same format - in this case pcds. If you've got a mixture, then you can always trim and clean the dataset in order to create a pcd version of the post code and link on that:

See some definitions below:

Three variations of formatting for each postcode are presented:

• pcd: 7-character version of the postcode (e.g. 'BT1 1AA', 'BT486PL')
• pcd2: 8-character version of the postcode (e.g. 'BT1  1AA', 'BT48 6PL')
• pcds: one space between the district and sector-unit part of the postcode (e.g. 'BT1 1AA', 'BT48 6PL') - possibly the most common formatting of postcodes.

Source: http://data.nicva.org/dataset/ons-postcode-products

Hopefully that's some help, but let me know if you have any problems.

[Anonymous · Accepted solution]

HI @Anonymous ,

I think you need to add an address mapping table(mapping address and long/lat) as the bridge to link postcode table that contains long/lat information, then use the mapping table to link to your user table address field.

Tips and Tricks for Power BI Map visualizations#in-the-dataset-tips-to-improve-the-underlying-dataset 

Relationship in Power BI with Multiple Columns 

Regards,

Xiaoxin Sheng

[Anonymous]

Thank you both for the info, I have managed to get my data linked to the Stats data, and now Im using the correct field it seems to be working well....thank you.

Does PowerBi not support the UK in terms of counties, towns etc as all guides i can find are based on zip code or states which look very easy to manipulate!?

Thanks again

[davegwynedd · Accepted solution]

I've been using the various post code files from National Statistics and there are usually a number of different post code columns in the files. I'm assuming you've transformed the data in some way?

I can see 3 post codes in the NSPL_AUG_2019_UK file - headers: pcd, pcd2 and pcds. We had an issue with using the first two on our datasets but matched pretty well on pcds. Should be the same on the Nov release.

This assumes your post code data is clean - and that it's all in the same format - in this case pcds. If you've got a mixture, then you can always trim and clean the dataset in order to create a pcd version of the post code and link on that:

See some definitions below:

Three variations of formatting for each postcode are presented:

• pcd: 7-character version of the postcode (e.g. 'BT1 1AA', 'BT486PL')
• pcd2: 8-character version of the postcode (e.g. 'BT1  1AA', 'BT48 6PL')
• pcds: one space between the district and sector-unit part of the postcode (e.g. 'BT1 1AA', 'BT48 6PL') - possibly the most common formatting of postcodes.

Source: http://data.nicva.org/dataset/ons-postcode-products

Hopefully that's some help, but let me know if you have any problems.