Our study shows the importance of drinking water monitoring using simple but powerful visualization tools to better understand spatial variations in water quality. The paper reports Chernoff's Faces visualiza­tion approach applied for the classification of drinking water samples collected at twelve various districts of Gdańsk (Poland), over the period 1993-2000. A good visualization should give the viewer a rapid understanding of the data and the phenomenon behind the data. The complex data matrix containig 1756 results of determination of disinfection by-products (THMs: CHCl₃, CHBrCl₂+C₂HCl₃, CHBr₂Cl, CHBr₃ and organohalogen compounds: CCl₄, CH₂Cl₂, C₂Cl₄, C₂H₃Cl₃) was successfully treated with Chernoff's approach, yielding two different groups of similarity among the sampling sites, and reflecting different types of drinking water supplies (surface and groundwater).