Lots of people ask the question "What should I upgrade?" or "Should I upgrade this or build that?". Here is the method I use to answer these types of questions. Warning: this post contains algebra.
Intro: You have two choices, upgrading building X or building Y. Building X will increase your income by $x per hour, building y will increase your income by $y per hour. Also, building X takes tx hours to upgrade and building Y takes ty hours.
First of all, assume that you're going to do both upgrades, the only question is which you're going to do first. When both upgrades are complete, your income will increase by $x+$y per hour. The choice you make only affects how much money you earn during the intermediate period before both buildings have been upgraded.
Choice #1 - upgrade X first: After tx hours, your income will increase by $x per hour. After tx+ty hours, your income will increase by $x+$y per hour. During the intermediate period, you'll earn ty*$x from building X.
Choice #2 - upgrade Y first: Same idea, during the intermediate period, you'll earn tx*$y from building Y.
So, the choice is simple. If ty*$x > tx*$y, pick building X first, otherwise pick building Y.
Wrinkle #1: tx and ty are not simply how long it takes to upgrade building X or Y. It also covers how long it takes to earn the money to pay for the upgrade. For expensive buildings like nightclubs, the time required to earn enough money could be much longer than the upgrade time. If that's the case, you should calculate tx as the earning time, not the upgrade time.
Wrinkle #2: Limited Time Building bonus prizes effectively reduce the cost of an upgrade. You should decide how much money you think a prize is worth. If the prize is worth more than the money you would lose from upgrading an LTB over another building, then the LTB is the right choice.
Wrinkle #3: When deciding whether or not to build an LTB, the choice is more complicated because you don't have the option of postponing the LTB for very long. If you build the LTB, you lose money because you postpone better upgrades, but if you don't build the LTB, it will disappear forever. In this case, I would calculate how much money you lose from building the LTB early and add that to the cost of the build. Then I would figure out how long it would take for the LTB to pay back the cost and decide if that period is too long. Example: building a $50 mil LTB before an NC upgrade costs $30 mil in lost NC revenue. So the LTB is effectively an $80 mil building. If the IPH is $40k, then it will take 2000 hours or about 3 months to pay back the total cost.