I’ve already done the first question till 4- the only thing I ask you to help is review the question 1 to 4 and complete the question 5 and 6.
Do your data analysis and answer the following questions for Baggage Complaints:
1) Which airline has the best record?
2) Which airline has the worst record?
3) Are complaints getting better or worse over time?
4) What other factors affect baggage performance?
5) Complete a time series analysis on Baggage complaint% by Airline. Find and report the best forecasting method for each airline. Report forecast value for baggage complaints next month (the first month after the last observation in your data set). Time series techniques:
Moving average
Weighted moving average
Exponential smoothing
Use MAE and MAPE to judge each forecasting technique
6. Step by Step Guide for Seasonality and Trends:
Seasonality:
Check out your group’s case5-Seasonal_Index_Included file.
For each airline, choose the best forecast method from Step 5 (Moving Average, Weighted Moving Average, Exponential Smoothing) and see if incorporating seasonality improves the method. The seasonal index is included in the case5 seasonal index file, so all you need to do is:
Deseasonalize the data
Run your best forecast method on that deseasonalized data (MA, WMA, or Exp. Smoothing). You can keep the same weights or alpha.
Readjust for seasonality using the Seasonal Index (provided)
Calculate the MAE and MAPE
Does your error percentage improve when using seasonality?
Trends:
Using a pivot table, having Airlines in the columns, Year in the rows, and Average Baggage% in the values, see the baggage% for each airline over time.
Insert a Pivot Line Chart or regular Line Chart to examine if there are any 3-7 year periods where ONE AIRLINE had a positive or negative trend.
Put that data for that airline over that time period in a new tab (United from 2013-2016 for example)
In two columns, put Month and Baggage%. Month 1 will be the start of your trend and keep counting up in that column (Jan 2013 = Month 1 in the above example and Dec 2016 = Month 48). Baggage% is in the column next to it and corresponds to the data for that time period for that airline.
Run a regression analysis under Data -> Data Analysis -> Regression and use the equation:
Forecast = y-intercept coefficient + Month coefficient * Month
Compare that forecast for each month to your Baggage% observation for those 3-7 years.
What is your MAE and MAPE using this trend to forecast?
Trend + Seasonality:
Using the same data subset from 7 (3-7 years for 1 airline), deaseasonalize your data.
Run a regression on the deseasonalized data (following same steps from 7, but this time on the deseasonalized data).
Readjust for seasonality using the Seasonal Index
Put the adjusted forecast for those years in a column next to the Baggage %
Calculate MAE and MAPE. Did your error percentage improve from Question 7 to 8?Do your data analysis and answer the following questions for Baggage Complaints:
Which airline has the best record?
Which airline has the worst record?
Are complaints getting better or worse over time?
What other factors affect baggage performance?
Complete a time series analysis on Baggage complaint% by Airline. Find and report the best forecasting method for each airline. Report forecast value for baggage complaints next month (the first month after the last observation in your data set). Time series techniques:
Moving average
Weighted moving average
Exponential smoothing
Use MAE and MAPE to judge each forecasting technique
6. Step by Step Guide for Seasonality and Trends:
Seasonality:
Check out your group’s case5-Seasonal_Index_Included file.
For each airline, choose the best forecast method from Step 5 (Moving Average, Weighted Moving Average, Exponential Smoothing) and see if incorporating seasonality improves the method. The seasonal index is included in the case5 seasonal index file, so all you need to do is:
Deseasonalize the data
Run your best forecast method on that deseasonalized data (MA, WMA, or Exp. Smoothing). You can keep the same weights or alpha.
Readjust for seasonality using the Seasonal Index (provided)
Calculate the MAE and MAPE
Does your error percentage improve when using seasonality?
Trends:
Using a pivot table, having Airlines in the columns, Year in the rows, and Average Baggage% in the values, see the baggage% for each airline over time.
Insert a Pivot Line Chart or regular Line Chart to examine if there are any 3-7 year periods where ONE AIRLINE had a positive or negative trend.
Put that data for that airline over that time period in a new tab (United from 2013-2016 for example)
In two columns, put Month and Baggage%. Month 1 will be the start of your trend and keep counting up in that column (Jan 2013 = Month 1 in the above example and Dec 2016 = Month 48). Baggage% is in the column next to it and corresponds to the data for that time period for that airline.
Run a regression analysis under Data -> Data Analysis -> Regression and use the equation:
Forecast = y-intercept coefficient + Month coefficient * Month
Compare that forecast for each month to your Baggage% observation for those 3-7 years.
What is your MAE and MAPE using this trend to forecast?
Trend + Seasonality:
Using the same data subset from 7 (3-7 years for 1 airline), deaseasonalize your data.
Run a regression on the deseasonalized data (following same steps from 7, but this time on the deseasonalized data).
Readjust for seasonality using the Seasonal Index
Put the adjusted forecast for those years in a column next to the Baggage %
Calculate MAE and MAPE. Did your error percentage improve from Question 7 to 8?
