Abstract
	Available software for the statistical data treatment offers a list of various point estimates of the location and spread parameters but rarely is able to help the user to choose the parameter statistically adequate to an actual sample batch. The exploratory data analysis and an examination of sample assumptions will find an answer to this question.  Rigorous chemometrical methodology presented on the metallurgical data concerning an estimation of the mean value and requested lower quantiles of the tensile strength of a construction steel illustrates the proposed procedure of a statistical data treatment with the exploratory data analysis. The sample distribution is skewed. Under such circumstances the original data should be transformed. The Box-Cox transformation improves a sample symmetry and also make a stabilization of a spread. The plot of logarithm of the likelihood function enables us to find an optimal transformation parameter. Proposed procedure gives reliable estimates of location parameter for asymmetric distribution when an arithmetic mean can not be used.