Discuss at least three factors a service industry such as an airline need to consider when forecasting demands. How do these factors affect operation decisions? For example, Taco Bell needs to consider the demand difference during lunch time and dinner time and use the data to schedule staff work hours.

The first factor is that services cannot be warehoused. In other words, if you are manufacturing paper towels and sell fewer than forecasted one week, you can warehouse the unsold inventory and sell it the next week. If a seat on an airplane is not filled, you lose money. An unsold seat on the 1 November 10 a.m. flight from JFK to LAX cannot somehow be stored and sold on 5 November.
Next, demand must be forecast precisely in terms of specific places and times. If you manufacture paper towels, it really does not matter if the customer buys them in store 1 or store 2, at 8 a.m. or 8 p.m.; you can just keep enough inventory on hand to supply both stores with a few weeks of the product. In the case of airplane seats, it is important to have the right number of seats available at precisely the right time. That means collecting data to discover the precise times at which people want to take flights to specific destinations and the profitability of different seats. For example, an 8 a.m. Monday flight from JFK to LAX might have more profitable business customers than a Saturday afternoon flight appealing to leisure travelers. 
Another extremely important part of forecasting in service industries is estimating the differences between peak and off-peak demands. Because many airline workers must have specific skills and training, one needs to plan how to deploy employees in a way that will cope with peak demand while not resulting in idle employees in slower periods. One cannot just hire high school kids to fly airplanes or serve as the cabin crew over the Christmas holidays.
https://hbr.org/1976/11/match-supply-and-demand-in-service-industries