The "floor" of x gives you the integer part obtained by rounding down, so e.g. floor(12345.67) = 12345. (I'm purposely sweeping the issue of recurring 9s in the decimal under the rug here.)
Assuming N is a nonnegative integer, the remainder of N when divided by 10 (often written N % 10) is the last digit, so e.g. 12345 % 10 = 5.
So the last digit before the decimal point of a nonnegative real number x is floor(x) % 10.
If x is negative then you can just flip the sign by taking the absolute value |x| and then apply the above, so the formula that works for all real numbers x is floor(|x|) % 10.