3. THE DEVELOPMENT OF THE MODEL 
Many entities with their corresponding attributes have 
been preliminary defined in the GIS. One of the most 
important considered entity, according to the main task of 
the GIS itself, is the ‘business activity’ and its attributes. 
SUP 
COD 
NTP 
FLU 
DCE 
DST 
DCG 
55 
5867 
9 
1 
354,5 
869,9 
151,3 
162 
7928 
9 
2 
225,1 
429,7 
176,7 
100 
7567 
9 
1 
378,2 
872,6 
284 
30 
8041 
9 
2 
469,8 
984,7 
200,3 
36 
4782 
1 
3 
485,4 
999,7 
216,6 
The variance analysis has been applied, in order to find 
the expression which shows the relations, among all the 
variables starting from the available data. 
In this case, the effects of the variables (attributes): 
• the type of the business activity (NTP); 
• the surface of the business activity (SUP); 
• its distance from the main street (DCG), 
• its distance from the station (DST); 
• the flow of clients for each activity (FLU). 
acquired in different ways and referred to the city of 
Reggio Calabria, have been studied, modeled, estimated 
and tested. 
In order to join the goal, firstly there have been done the 
plots and the summarizing of data and then it has been 
built up the suitable model, step by step. 
From the first analysis, the most influent variables on the 
flow of clients have been found. Consequentially only 
those have been considered, for the implementation of 
the expected model. Then the relations, among these 
variables, both separately for each type of activity and 
globally all together, have been found 
The different models, for each type of business activity, 
obtained by non - linear regressions and variance 
analysis, are shown in the following tables, where the 
values of parameters, standard errors, correlation 
coefficients, etc. are summarized. 
• For type ‘ABB’ (clothes, shoes, etc): 
*** Non - linear Regression Model *** 
Formula: FLU ~ B0/(DCG A 2 + Bl) + B2 * SUP 
Parameters: 
Value 
Std. Error 
t value 
B0 
2.73394e+005 
8.60574e+004 
3.17688 
Bl 
3.08994e+004 
9.35960e+003 
3.30136 
B2 
3.62716e-002 7 
40886e-003 
4.89570 
Residual standard error: 0.782917 
on 7 degrees of freedom 
Correlation of Parameter Estimates: 
BO Bl 
Bl 0.965 
B2 -0.460 -0.266 
• For the type ‘ALIM’ (food): 
*** Non - linear Regression Model *** 
Formula: FLU ~ B0 + Bl * sin(DCG) + B2 * sin(SUP) 
Parameters: 
Value 
Std. Error 
t value 
B0 
4.74389 
0.833059 
5.69454 
Bl 
-2.59448 
0.721481 
-3.59604 
B2 
-3.29219 
1.486830 
-2.21423 
Residual standard error: 0.843604 
on 3 degrees of freedom 
Correlation of Parameter Estimates: 
B0 Bl 
Bl 0.668 
B2 0.910 0.709 
• For all types: 
*** Non - linear Regression Model *** 
Formula: 
FLU ~ V8 * (B0/(DCG A 2 + Bl) + B2 * SUP) + 
+(1 - V8) * (B3 + B4 * sin(DCG) + B5 * sin(SUP)) 
Parameters: 
Value 
Std. Error 
t value 
B0 
2.73394e+005 
8.81116e+004 
3.10281 
Bl 
3.08994e+004 
9.58302e+003 
3.22439 
B2 
3.62716e-002 
7.58571e-003 
4.78156 
B3 
4.743 89e+000 
7.91586e-001 
5.99289 
B4 
-2.59448e+000 
6.85563e-001 
-3.78445 
B5 
-3.29219e+000 
1.41281e+000 
-2.33023 
Residual standard error: 0.801606 
on 10 degrees of freedom 
Correlation of Parameter Estimates: 
B0 
Bl 
B2 
B3 
B4 
Bl 
0.965 
B2 
-0.460 
-0.266 
B3 
0.000 
0.000 
0.000 
B4 
0.000 
0.000 
0.000 
0.668 
B5 
0.000 
0.000 
0.000 
0.910 
0.709