This study discusses the problem of modeling inherently positive data sets using Modified Quadratic Shepard interpolation method. There are number of scientific and business domains where data is inherently positive. For example numbers of people, mass, volume and percentage mass concentration are meaningless when negative. This interplant generates negative values while modeling such data that may cause ambiguity. We present a scaling technique that constrains the interpolant to produce non-negative graph through scattered positive data sets.