Experimentation with the addition, omission and substitution of variables. Which variables ARE most important. What impact does the addition/subtraction/substitution of variables (such as commuting or distance) have on the cluster output?

• Experimentation with the addition, omission and substitution of variables. Which variables ARE most important. What impact does the addition/subtraction/substitution of variables (such as commuting or distance) have on the cluster output?

• Does 8 clusters remain the optimum solution?

• Experimentation with different measures of proximity and methods of standardisation if necessary.

• Optimisation of the classification through the use of an optimising clustering algorithm such as k-means.

• A different classification based on migrant flows rather than migrants? Wards rather than districts?