Molecular Modeling Pro Plus 2-D Graph Output


Below is an example of the options panel for the X-Y plot (figure 16) and the resulting plot.  It is taken from a tutorial in the on-line help file.  This tutorial uses the development of a model for flash point as an example.



Figure 16.  The options panel for the "advanced" 2-D XY plot.  We are going to color the plot by a third variable (connectivity index 1).  The minimum and maximum x and y values have been rounded for better looking graph labels.  We also are requesting that the x axis undergo the 1/x transform.


Hit the done button in the options panel.  The list of fields again appears so you can select the field to color the data points by.  Select connectivity_1 from the list.  The plot in figure 17 appears.


Figure 17.  Example of the XY (2-D) "advanced" plot.  Flash point is fairly well correlated with 1/enthalpy of vaporization (the curve drawn through the data points is that made by the linear least squares regression model).  Data points are colored by a third variable (connectivity index 1).  It appears that enthalpy of vaporization and connectivity index 1 are probably also intercorrelated as the colors of the data points are clustered.  The model flash point = 1/enthalpy of vaporization accounts for about 75% of the variance for 360 values of flash point of solvents and surfactants found in the industrial.mdb database.  Most of the values with lower connectivity indices (dark blue) are solvents and are found toward the left side of the graph.  Their behavior is different from surfactants.  The Joback and Reid enthalpy of vaporization calculation was built with solvents well-represented and surfactants not represented.  A question to ask about this model is that is the inverse transform really needed or is it reflecting inaccuracy in the calculation of enthalpy of vaporization.  Note that a very straightforward linear relationship appears to exist between flash point and enthalpy of vaporization for solvents.  These sorts of insights can only be found by plotting the data.  Note you can easily see out-lying data points in this plot.