This idea once had my un-divided attention.

STADT starts with mapping a city as it is now, every business or residental lot. The business are carefully categorized, with the goal of placing competing business in the same group.

Let's focus on a single business: drugstores, for example. 

A complex formula would calculate influence of a drugstore on a certain plot. It would translate the idea of having a drugstore nearby or not. This formula would be made up, but would take into account the area of a drugstore (m2) and the distance from the drugstore to the place. The residential plots, when nearby, would give a bonus to the formula (I'll explain why in a minute).

Summing over all drugstores in the city, you would have the presence of drugstore services in that place.

The idea is to repeat this calculation all over the city or at least the lots you know to be for rent or sale.

What this would indicate is the best places in town to open a new drugstore. This works best for business that are more local, where people are not willing to travel too far for them. Hypermarkets, for example, work on much greater distance than grocery stores.

The residential plots would give a bonus to the formula because they would make the place attractive for new business, because of the customers. They would serve as an attractor. In fact, when I first worked on this, I separated low density residential zones (houses), medium density (small buildings) and high density (skyscrapers).

The company based on such ideas would sell the consulting services: people wanting to open business would seek them out for the information.

It is a great idea, but it is far from being a small one. At first I considered raising information about the lots of the city using Google Maps, but I soon found out it is an explicit violation to their terms of service. There is not such thing as a free meal.