It might be more do-able to consider a calculation based on Lat & Long to determine closest 1, closest 2, rather than the data itself.
In other words, use Lat Long for Office 2 in relation to Office 1 to confirm that mathematically it is in-fact the closest. You need the Spherical Law of Cosines.
Luckily, we have a KB for this. Now this example calculates from a point (Lat Long) to the nearest zip code. But you'd simply need to modify to measure from Point A (Office 1 Lat Long) to Point B (office 2), Point C, (office 3), and so on..
If you've found this post helpful and/or it's answered your question, please mark it as so, so that others can find it quickly.
Tableau Online Deployment Advisor