(Get Answer)

Assume that the data in columns A and B report the number of residential home starts, by month, in thousands, from 2015 to 2020 in the state of Oregon. Months Housing starts (000) Jan-2015 29 Feb-2015 32 Mar-2015 41 Apr-2015 44 May-2015 49 Jun-2015 47 Jul-2015 46 Aug-2015 47 Sep-2015 43 Oct-2015 45 Nov-2015 34 Dec-2015 31 Jan-2016 35 Feb-2016 43 Mar-2016 46 Apr-2016 46 May-2016 43 Jun-2016 41 Jul-2016 44 Aug-2016 47 Sep-2016 41 Oct-2016 40 Nov-2016 32 Dec-2016 32 Jan-2017 34 Feb-2017 40 Mar-2017 43 Apr-2017 42 May-2017 43 Jun-2017 44 Jul-2017 39 Aug-2017 40 Sep-2017 33 Oct-2017 32 Nov-2017 31 Dec-2017 28 Jan-2018 34 Feb-2018 29 Mar-2018 36 Apr-2018 42 May-2018 43 Jun-2018 44 Jul-2018 44 Aug-2018 48 Sep-2018 45 Oct-2018 44 Nov-2018 40 Dec-2018 37 Jan-2019 45 Feb-2019 49 Mar-2019 62 Apr-2019 62 May-2019 58 Jun-2019 59 Jul-2019 64 Aug-2019 62 Sep-2019 50 Oct-2019 52 Nov-2019 50 Dec-2019 44 Jan-2020 51 Feb-2020 56 Mar-2020 60 Apr-2020 65 May-2020 64 Jun-2020 63 Jul-2020 63 Aug-2020 72 Sep-2020 61 Oct-2020 65 Nov-2020 51 Dec-2020 47 Jan-2021 Feb-2021 Mar-2021 Apr-2021 May-2021 Jun-2021 Jul-2021 Aug-2021 Sep-2021 Oct-2021 Nov-2021 Dec-2021 a) Plot the data and specify any evidence of trend and/or seasonality. b) Use the "Tahoe salt" method with linear trend to forecast housing starts for the 12 months of 2021 (1 decimal place). c) Overlay the forecasts on your graph with the data. It is not necessary to overlay the intermediate steps (disaggregation, linear projection) but they may be useful as you develop the model. d) Calculate the model's MAE (1 decimal place) e) Calculate and plot the drift signal in a separate graph. f) Analyze the quality of this model. What is the glaring defect? What is the main cause of this defect? g) Suggest an approach that would give better results in terms of prediction values, MAE and drift signal. An appropriate and justified narrative argument is sufficient (so one or two sentences). No calculations are required for this section.