What is happiness? That's a question everyone can answer differently. Some people are happy, when they win a race, someone else is happy he survived cancer etc. We'll try to define happiness as a feeling.
If we look at this problem from geek's point of view, we'll find out, that happiness can be described as magnitude. Let's mark it H and we will try to find it's numeric value.
If we want to do that, we need to specify borders of that value. I took inspiration from Mr. Celsius and his temperature scale. I've choosen absolute happiness and absolute bad luck as borders. That leads us to the best base value - %.
If we are feeling happy depends on many factors. The biggest effect will probably have our common view of the world. Admit it – If you aren't trying to find happy sites of stuff, you won't be happy. Let's call this character factor and mark it F_c. It will be defined in range from 0 to 50, with 0 being most pesimistic and 50 most optimistic value. This value is changing in long time, but it is individual and variable nonetheless, so it should be measured before each calculation, by series of tests. Next significant factor will be current mood - F_m. It is defined in range from 0 to 25. Again an example: You'll find two bucks just lying on the floor. Normally you would be at least a little bit happy, but if you just had a dispute with your boss, two bucks probably won't be enough to make you happy. the influence of remaining factors is insignificant by comparasm, but there is too many of them, like your wealth, level of your social life etc. Again, most of them are individual factors and hard to describe, so we connect them to one factor, satisfaction - S, again ranging from 0 to 25. Now we can make the happiness formula:
H = F_c + F_m + S; [H] = %
Like everywhere else in the nature, there will be some exceptions. First case applies, when the subject already died. If we want to calculate happiness of a dead person, we must know data from time prior to the subject's demise. BUT! In that case, ve can't define mood anymore, so we are going to leave it out Good. Now we'll try to insert model data. Imagine you are a pesimistic soldier, who lost all limbs in Vietnam. Euthanasia is forbidden and he is sufferning for the rest of his life, let's say 75 years, until he finally dies and is freed from his misery:
H = F_c + S = 0 + 0 = 0%
Now your mind is probably asking what the hell? Would you consider your death in this case a bad luck? I sure wouldn't. That indicates we need to modify the formula again, to catch the inverted value:
H = 100 – (F_c + S)
Insert our model data:
H = 100 – (0 + 0) = 100%
that's better. Still, situation in wich his death is complete loss isn't possible, even if he was satisfied optimist, but think about it. What are you doing every day? You are always stressed. You need to pay your bills, go to work, take care of children... He is now stripped of these duties.
Next problem can be getting the data at first. Imagine getting valid data from newborns and people with some mental disease.
Now we have enough materia to define basic axioms of happiness:
H can not be calculated for people incapable of providing valid data.
H can not be measured directly, we are using three helping factors - Character factor F_c, Mood factor F_m and Satisfaction S.
H = Kc + Pa + Sp; [H] = % for living subjects.
H = 100 – (Kc + Sp); [H] = % for deceased subjects.

Our search indicates, that happiness is, in specific frame, measurable. We can't define specific reasons for happiness, we can only know if we are happy and how much.