I assume cities are part of regions and suburbs are part of cities. If so, you may need more tables.
One table for your regions, which has all the information for them with a column ID (auto-incrementing).
One table for your cities, which has all the information for them with a column ID (auto-incrementing).
One table for your suburbs, which has all the information for them with a column ID (auto-incrementing).
And then, one table relating your cities to your regions. One column is ID (auto-increment), one is city (where your city ID's from your city table go), and one is region (where you region ID's from your region table go). This lets you know which cities belong to which region.
Next, one table relating your suburbs to your cities. One column is ID (auto-increment), one is region (once again from your region table), and one is suburb (from your suburbs table). This lets you know which suburbs belong to which cities.
With this setup, you can use INNER JOIN to get your cities or suburbs and you will have both the data for the suburb and the data for the more-general categories.
Hope this helps, I've been developing relational web applications for a long time, and have found that my aforementioned method works very well.