One usually plots spectral density of the intensity vs. wavelength. However, if you try to plot it vs. frequency, you might find some interesting reslut. Make sure you use the transformation for the Jacobian as well:
<br />
I_{\lambda}(\lambda) \, |d\lambda| = I_{\nu}(\nu) \, |d\nu|, \; \lambda = c/\nu<br />
<br />
I_{\nu}(\nu) = I_{\lambda}\left(\frac{c}{\nu}\right) \, \left|\frac{d \lambda}{d \nu}\right| = \frac{c}{\nu^{2}} \, I_{\lambda}\left(\frac{c}{\nu}\right)<br />
I suggest you take several points from an experimental spectrum, change the ordinate and use the above Jacobian transformation and plot it.