In two recent publications we used a million and a half extragalactic
spectra from the Sloan Digital Sky Survey (SDSS) to study the diffuse
interstellar bands (DIBs) in the Milky Way. These weak absorption
lines, coming from the interstellar medium, have been known for a
century but their carriers evade us still. Although a single SDSS
spectrum lacks the necessary signal to noise to detect DIBs, we found
that stacking thousands of spectra allows their detection. We made the
first statistical measurement of DIB behaviour at high Galactic
latitude, mapped their distribution on the sky, and measured the
correlation of 142 DIBs with dust. We further grouped them using a
Machine Learning algorithm we developed for the purpose. We found 6
groups of DIBs that may come from the same carrier, 4 of these groups
are tightly associated with C2 and CN molecules. I will discuss these
results as well as some ideas for future research.