Figure 1 represents the location of the different polluted sites in the commune of Vernier in comparison with DCGP values for each points corresponding to residential areas. We can deduce a spatial correlation between those parameters only by observing the map. Indeed, points in darker red, corresponding to high DCGP value, tends to be located closer to polluted site than points with low DCGP values. This spatial correlation could be seen at locations A,B and C (in green) for instance.
Second step was to confirm this spatial correlation with a spatial data analysis. The distance matrix created in QGIS has therefore been exported to Geoda in the purpose of creating a scatterplot between the distance to a polluted site and the distance to the closest generalist practitioner.
As the obtained scatterplot highlights, a spatial correlation exists between the mean distance to a polluted site (on y-axis) and the distance to the closest generalist practitioner. Indeed, the regression line has a positive slope closed to 1 (b=0,656) meaning that the two parameters are positively correlated. This is not the result expected. People living far from a polluted site are supposed to be the closest to a generalist practitioner. Figure 1 confirm this assumption. The scatterplot however highlights the fact that distance to polluted sites increases when ditance to closest generalist practiotioner increases.