Expanding on LDC3's comment a little.
The information in the table (irrelevant of how good or bad the method proposed is) tells you what the saturation points are for solvent mixtures at given temperatures for each compound.
Fundamentally, you want to separate your desired product from your undesired product. In this case, you are trying to recrystallise the 3-phenylacrylaldehyde, leaving the benzaldehyde in solution. You know this because for all temperatures and solvent mixtures it has lower solubility (this cannot be assumed for a more complex solvent array).
You need the minimum of hot solvent, here 100 degrees, to dissolve all of your crude compound, and you want to maximise the return of pure compound on cooling. This explains why you want the biggest ratio of hot to cold solubility as LDC3 indicated.
You also want to minimise the amount of your starting material that crashes out with your product. So, calculate the same ratios for the starting material, but here you want to minimise the return on cooling, so you want the smallest ratio.
You now have a table of solubility ratios, columns being 'solvent mix', 'rP' (ratio of product solubilities) and 'rSM' (ratio of starting material solubilities). The largest value of rP/rSM gives you the best solvent mixture.