Northcutt Bikes Answers2
Northcutt Bikes Case
Answers
1
Q1: Demand Data Plot
2
Q1: Plot Shows
There is seasonality
There is a trend
Forecast should take into account both
3
Construction of base indices
Year:
January
February
March
April
May
June
July
August
September
October
November
December
Mean Demand:
2008
0.53
0.74
0.88
1.09
1.10
1.60
1.29
1.19
1.00
1.09
0.73
0.74
2009
0.72
0.74
0.84
1.00
1.16
1.57
0.94
1.30
1.13
0.74
0.99
0.88
818.42
990.50
2010
0.59
0.95
0.79
1.18
1.15
1.39
1.35
1.43
0.91
0.96
0.78
0.51
2011
0.59
1.09
0.98
0.92
1.27
1.51
1.56
0.71
1.08
0.77
0.84
0.67
Mean Base
0.61
0.88
0.87
1.05
1.17
1.52
1.28
1.16
1.03
0.89
0.84
0.70
1032.08 1181.25
4
Multiple Regression Results:
X is Period and Base
Regression Statistics
Multiple R
0.982917071
R Square
0.966125969
Adjusted R Square 0.964620456
Standard Error
59.82147676
Observations
48
ANOVA Regression
Residual
Total Intercept
Period
Base
df
2
45
47
SS
MS
F
4592970.404 2296485.202 641.7256395
161037.4087 3578.609082
4754007.813
Coefficients Standard Error t Stat
P-value
-219.4209094
35.31667659 -6.212954633 1.50687E-07
8.730540524
0.623285303 14.00729407 5.12015E-18
1011.295853
30.74315604 32.89499139 4.07081E-33
5
Q2: Forecasting Methods
Multiple regression or MR (Y is forecast,
X’s are period and base) MAD ≈ 45.096
Simple regression or SR (deseasonalize demand, seasonalize forecast, X is period) MAD ≈ 32.403
Exponential Smoothing or ES (adjusted for trend and seasonality) MAD ≈ 13.258
6
Q2: Forecast for
January – April 2012
Month
Mean Base Period
MR
SR
ES
January
0.61
49
825.27
745.12
720.56
February
0.88
50
1107.05 1082.68
1039.50
March
0.87
51
1105.66 1078.04
1027.69
April
1.05
52
1296.43 1310.32
1240.31
7
Q3: Best Forecast:
Exponential smoothing forecast has lowest MAD
Disadvantages: the exponential smoothing forecast should be updated frequently (say once a month).
8
Q4: Additional Information
Jan’s knowledge of market could be
Answers
1
Q1: Demand Data Plot
2
Q1: Plot Shows
There is seasonality
There is a trend
Forecast should take into account both
3
Construction of base indices
Year:
January
February
March
April
May
June
July
August
September
October
November
December
Mean Demand:
2008
0.53
0.74
0.88
1.09
1.10
1.60
1.29
1.19
1.00
1.09
0.73
0.74
2009
0.72
0.74
0.84
1.00
1.16
1.57
0.94
1.30
1.13
0.74
0.99
0.88
818.42
990.50
2010
0.59
0.95
0.79
1.18
1.15
1.39
1.35
1.43
0.91
0.96
0.78
0.51
2011
0.59
1.09
0.98
0.92
1.27
1.51
1.56
0.71
1.08
0.77
0.84
0.67
Mean Base
0.61
0.88
0.87
1.05
1.17
1.52
1.28
1.16
1.03
0.89
0.84
0.70
1032.08 1181.25
4
Multiple Regression Results:
X is Period and Base
Regression Statistics
Multiple R
0.982917071
R Square
0.966125969
Adjusted R Square 0.964620456
Standard Error
59.82147676
Observations
48
ANOVA Regression
Residual
Total Intercept
Period
Base
df
2
45
47
SS
MS
F
4592970.404 2296485.202 641.7256395
161037.4087 3578.609082
4754007.813
Coefficients Standard Error t Stat
P-value
-219.4209094
35.31667659 -6.212954633 1.50687E-07
8.730540524
0.623285303 14.00729407 5.12015E-18
1011.295853
30.74315604 32.89499139 4.07081E-33
5
Q2: Forecasting Methods
Multiple regression or MR (Y is forecast,
X’s are period and base) MAD ≈ 45.096
Simple regression or SR (deseasonalize demand, seasonalize forecast, X is period) MAD ≈ 32.403
Exponential Smoothing or ES (adjusted for trend and seasonality) MAD ≈ 13.258
6
Q2: Forecast for
January – April 2012
Month
Mean Base Period
MR
SR
ES
January
0.61
49
825.27
745.12
720.56
February
0.88
50
1107.05 1082.68
1039.50
March
0.87
51
1105.66 1078.04
1027.69
April
1.05
52
1296.43 1310.32
1240.31
7
Q3: Best Forecast:
Exponential smoothing forecast has lowest MAD
Disadvantages: the exponential smoothing forecast should be updated frequently (say once a month).
8
Q4: Additional Information
Jan’s knowledge of market could be