Rose Sparkling Wine

For this particular assignment, the data of different types of wine sales in the 20th century is to be
analysed. Both of these data are from the same company but of different wines. As an analyst in the
ABC Estate Wines, you are tasked to analyse and forecast Wine Sales in the 20th century.
1.Read the data as an appropriate Time Series data and plot the data.
Rose wine sales
Sparkling -wine sale
2. Perform appropriate Exploratory Data Analysis to understand the data and also perform
decomposition.
Rose wine sales

Data set has 187 records and two null values. Null values have been imputed
Decomposition of Rose-wine sales into Trend, Seasonal and Residual
Distribution of sale of wine-Rose in each year

Distribution of sale of wine-Rose in each Month
Trend for each month over the years in sale of wine-Rose

Sparkling wine sales
Data set has 187 records and there are no null values in data set Sparkling
Decomposition of Sparkling-wine sales into Trend, Seasonal and Residual
Distribution of sale of wine-Sparkling in each year
Distribution of sale of wine-Sparkling in each Month

Trend for each month over the years in sale of wine-Sparkling
3. Split the data into training and test. The test data should start in 1991.
After splitting train and test data set , train data set has 132 records and test has 55 records
4. Build various exponential smoothing models on the training data and evaluate the model using RMSE
on the test data. Other models such as regression,naïve forecast models, simple average models etc.
should also be built on the training data and check the performance on the test data using RMSE.
Please do try to build as many models as possible and as many iterations of models as possible with different
parameters.
Various models and forecast as below.
Rose wine sales

Various test results and RMSE are as under.
Test RMSE
RegressionOnTime 15.280000
NaiveModel 79.741326
SimpleAverageModel 53.483727
2pointTrailingMovingAverage 11.529811
Alpha=0.995,SimpleExponentialSmoothing 36.819844
Alpha=0.995,Beta=0.995:DoubleExponentialSmoothing 15.276679
Alpha=0.99,Beta=0.0001,Gamma=0.005:DoubleExponentialSmoothing 20.962011

RMSE for models are as below.
Test RMSE
Test RMSE
5. Check for the stationarity of the data on which the model is being built on using appropriate statistical
tests and also mention the hypothesis for the statistical test. If the data is found to be non-stationary,
take appropriate steps to make it stationary. Check the new data for stationarity and comment. Note:
Stationarity should be checked at alpha = 0.05.
To check whether the series is stationary, we use the Augmented Dickey Fuller (ADF)test whose null and
alternate hypothesis can be simplified to
Null Hypothesis H0: Time Series is non-stationary
Alternate Hypothesis Ha: Time Series is stationary
Rose wine sales
Results of Dickey-Fuller Test:

Test Statistic -1.873307
p-value 0.344721
#Lags Used 13.000000
Number of Observations Used 173.000000
Critical Value (1%) -3.468726
dtype: float64
since p-value> 0.05, at alpha 0.05, time series is not stationary.

We can take next levels of differencing to make a Time Series stationary.
Test Statistic -8.044136e+00
p-value 1.813615e-12
#Lags Used 1.200000e+01
Number of Observations Used 1.730000e+02
Critical Value (1%) -3.468726e+00
dtype: float64
After next level of levels of differencing p-value <0.05 therefore series is stationary.

p-value 0.601061
dtype: float64
since p-value> 0.05, at alpha 0.05, time series is not stationary.
We can take next levels of differencing to make a Time Series stationary.

p-value 0.000000
dtype: float64
After next level of levels of differencing p-value <0.05 therefore series is stationary.
6. Build an automated version of the ARIMA/SARIMA model in which the parameters are selected using
the lowest Akaike Information Criteria (AIC) on the training data and evaluate this model on the test data
using RMSE.
Rose wine sales
ARIMA Model
AIC values in descending order
param AIC
17 (3, 1, 3) 1273.194108
4 (1, 1, 2) 1277.359223
3 (1, 1, 1) 1277.775747
param AIC
9 (2, 1, 1) 1279.045689
10 (2, 1, 2) 1279.298694
5 (1, 1, 3) 1279.312635
15 (3, 1, 1) 1279.605966
16 (3, 1, 2) 1280.969245
11 (2, 1, 3) 1281.196226
1 (1, 0, 2) 1292.053210
7 (2, 0, 2) 1292.248055
2 (1, 0, 3) 1292.929011
6 (2, 0, 1) 1292.937195
14 (3, 0, 3) 1293.042709
8 (2, 0, 3) 1294.247938
0 (1, 0, 1) 1294.510585
12 (3, 0, 1) 1333.933193
13 (3, 0, 2) 1355.403813
Lowest AIC is: 1273.194 with param (3,1,3)
ARIMA Model Results

==============================================================================
Dep. Variable: D.Rose No. Observations: 131
Model: ARIMA(3, 1, 3) Log Likelihood -628.597
Method: css-mle S.D. of innovations 28.355
Date: Sat, 06 Mar 2021 AIC 1273.194
Time: 14:46:16 BIC 1296.196
Sample: 02-01-1980 HQIC 1282.541
- 12-01-1990
================================================================================
coef std err z P>|z| [0.025 0.975]
--------------------------------------------------------------------------------
const -0.4906 0.088 -5.548 0.000 -0.664 -0.317
ar.L1.D.Rose -0.7244 0.086 -8.417 0.000 -0.893 -0.556
ar.L2.D.Rose -0.7218 0.086 -8.349 0.000 -0.891 -0.552
ar.L3.D.Rose 0.2763 0.085 3.236 0.001 0.109 0.444
ma.L1.D.Rose -0.0150 0.044 -0.338 0.735 -0.102 0.072
ma.L2.D.Rose 0.0150 0.044 0.339 0.734 -0.072 0.102
ma.L3.D.Rose -1.0000 0.046 -21.918 0.000 -1.089 -0.911
Roots
=============================================================================
Real Imaginary Modulus Frequency
-----------------------------------------------------------------------------
AR.1 -0.5011 -0.8661j 1.0006 -0.3335
AR.2 -0.5011 +0.8661j 1.0006 0.3335
AR.3 3.6147 -0.0000j 3.6147 -0.0000
MA.1 1.0000 -0.0000j 1.0000 -0.0000
MA.2 -0.4925 -0.8703j 1.0000 -0.3320
MA.3 -0.4925 +0.8703j 1.0000 0.3320
-----------------------------------------------------------------------------
Evaluation of model using RMSE: RMSE from ARIMA model 15.99
SARIMA model
AIC values in descending order (lowest AIC 10 records)
param seasonal AIC
107 (0, 1, 2) (2, 1, 2, 12) 774.969119
215 (1, 1, 2) (2, 1, 2, 12) 776.940114
323 (2, 1, 2) (2, 1, 2, 12) 776.996102
269 (2, 0, 2) (2, 1, 2, 12) 780.716942
161 (1, 0, 2) (2, 1, 2, 12) 780.992971
89 (0, 1, 1) (2, 1, 2, 12) 782.153872
322 (2, 1, 2) (2, 1, 1, 12) 783.703652
197 (1, 1, 1) (2, 1, 2, 12) 783.899095

param seasonal AIC
95 (0, 1, 2) (0, 1, 2, 12) 784.014096
311 (2, 1, 2) (0, 1, 2, 12) 784.140949
Lowest AIC is : 774.969 with param (0,1,2) and seasonal (2,1,2,12)
SARIMAX Results
===================================================================================
Dep. Variable: y No. Observations: 132
Model: ARIMAX(0, 1, 2)x(2, 1, 2, 12) Log Likelihood -380.485
Date: Sat, 06 Mar 2021 AIC 774.969
Time: 13:41:44 BIC 792.622
Sample: 0 HQIC 782.094
- 132
Covariance Type: opg
==============================================================================
coef std err z P>|z| [0.025 0.975]
------------------------------------------------------------------------------
ma.L1 -0.9524 0.184 -5.166 0.000 -1.314 -0.591
ma.L2 -0.0764 0.126 -0.605 0.545 -0.324 0.171
ar.S.L12 0.0480 0.177 0.271 0.786 -0.299 0.395
ar.S.L24 -0.0419 0.028 -1.513 0.130 -0.096 0.012
ma.S.L12 -0.7526 0.301 -2.503 0.012 -1.342 -0.163
ma.S.L24 -0.0721 0.204 -0.354 0.723 -0.471 0.327
sigma2 187.8679 45.274 4.150 0.000 99.132 276.604
===================================================================================
Ljung-Box (L1) (Q): 0.06 Jarque-Bera (JB): 4.86
Prob(Q): 0.81 Prob(JB): 0.09
Heteroskedasticity (H): 0.91 Skew: 0.41
Prob(H) (two-sided): 0.79 Kurtosis: 3.77
===================================================================================
Evaluation of model using RMSE : RMSE from SARIMA model 16.52
ARIMA Model
AIC values in descending order
param AIC
8 (2, 1, 2) 2210.616954
7 (2, 1, 1) 2232.360490
param AIC
2 (0, 1, 2) 2232.783098
5 (1, 1, 2) 2233.597647
4 (1, 1, 1) 2235.013945
6 (2, 1, 0) 2262.035601
1 (0, 1, 1) 2264.906439
3 (1, 1, 0) 2268.528061
0 (0, 1, 0) 2269.582796
Lowest AIC is 2232.783 with param (2,1,2)
ARIMA Model Results

==============================================================================
Dep. Variable: D.Sparkling No. Observations: 131
Date: Sun, 28 Feb 2021 AIC 2210.617
Time: 18:02:49 BIC 2227.868
Sample: 02-01-1980 HQIC 2217.627
- 12-01-1990
=====================================================================================
coef std err z P>|z| [0.025 0.975]
-------------------------------------------------------------------------------------
const 5.5860 0.516 10.825 0.000 4.575 6.597
ar.L1.D.Sparkling 1.2698 0.074 17.045 0.000 1.124 1.416
ar.L2.D.Sparkling -0.5601 0.074 -7.617 0.000 -0.704 -0.416
ma.L1.D.Sparkling -1.9993 0.042 -47.149 0.000 -2.082 -1.916
ma.L2.D.Sparkling 0.9993 0.042 23.584 0.000 0.916 1.082
Roots
=============================================================================
-----------------------------------------------------------------------------
AR.1 1.1335 -0.7075j 1.3361 -0.0888
AR.2 1.1335 +0.7075j 1.3361 0.0888
MA.1 1.0002 +0.0000j 1.0002 0.0000
MA.2 1.0006 +0.0000j 1.0006 0.0000
----------------------------------------------------
Evaluation of model using RMSE: RBSE for ARIMA model 1375.03

SARIMA model
AIC values in descending order (lowest AIC 10 records)
param seasonal AIC
95 (1, 1, 2) (0, 1, 2, 12) 1382.347780
41 (0, 1, 2) (0, 1, 2, 12) 1382.484254
101 (1, 1, 2) (1, 1, 2, 12) 1384.137874
149 (2, 1, 2) (0, 1, 2, 12) 1384.317618
47 (0, 1, 2) (1, 1, 2, 12) 1384.398867
107 (1, 1, 2) (2, 1, 2, 12) 1385.688721
53 (0, 1, 2) (2, 1, 2, 12) 1386.023734
155 (2, 1, 2) (1, 1, 2, 12) 1386.097242
161 (2, 1, 2) (2, 1, 2, 12) 1387.627785
77 (1, 1, 1) (0, 1, 2, 12) 1398.756167
Lowest AIC is : 1382.347780 with param (1,1,2) and seasonal 0,1,2,12)
SARIMAX Results
==================================================================================
Model: SARIMAX(0, 1, 2)x(0, 1, 2, 12) Log Likelihood -686.242
Date: Sun, 28 Feb 2021 AIC 1382.484
Time: 18:06:38 BIC 1395.093
- 132
==============================================================================
coef std err z P>|z| [0.025 0.975]
------------------------------------------------------------------------------
ma.L1 -0.7223 0.107 -6.752 0.000 -0.932 -0.513
ma.L2 -0.2292 0.137 -1.671 0.095 -0.498 0.040
ma.S.L12 -0.4113 0.087 -4.743 0.000 -0.581 -0.241
ma.S.L24 -0.0419 0.138 -0.304 0.761 -0.312 0.228
sigma2 1.736e+05 2.06e+04 8.425 0.000 1.33e+05 2.14e+05
===================================================================================
Prob(Q): 0.88 Prob(JB): 0.00
===================================================================================
Evaluation of model using RMSE: RBSE for ARIMA model 321.48
7. Build ARIMA/SARIMA models based on the cut-off points of ACF and PACF on the training data and
evaluate this model on the test data using RMSE.
Rose wine sales
ACF graph for Rose wine sales as under
PACF graph for Rose wine sales as under

It can be observed from above the ACF and PACF plots, the cut off points for p and q for ARIMA model is 4 and 2
respectively.
By taking these parameters (4,1,2) ARIMA results are as under
ARIMA Model Results

==============================================================================
Dep. Variable: D.Rose No. Observations: 131
Date: Sun, 07 Mar 2021 AIC 1283.753
Time: 13:23:57 BIC 1306.754
Sample: 02-01-1980 HQIC 1293.099
- 12-01-1990
================================================================================
coef std err z P>|z| [0.025 0.975]
--------------------------------------------------------------------------------
const -0.1905 0.576 -0.331 0.741 -1.319 0.938
ar.L1.D.Rose 1.1685 0.087 13.391 0.000 0.997 1.340
ar.L2.D.Rose -0.3562 0.132 -2.693 0.007 -0.616 -0.097
ar.L3.D.Rose 0.1855 0.132 1.402 0.161 -0.074 0.445
ar.L4.D.Rose -0.2227 0.091 -2.443 0.015 -0.401 -0.044
ma.L1.D.Rose -1.9506 nan nan nan nan nan
ma.L2.D.Rose 1.0000 nan nan nan nan nan
Roots
=============================================================================
-----------------------------------------------------------------------------
AR.1 1.1027 -0.4116j 1.1770 -0.0569
AR.2 1.1027 +0.4116j 1.1770 0.0569
AR.3 -0.6863 -1.6643j 1.8003 -0.3122
AR.4 -0.6863 +1.6643j 1.8003 0.3122
MA.1 0.9753 -0.2209j 1.0000 -0.0355
MA.2 0.9753 +0.2209j 1.0000 0.0355
-----------------------------------------------------------------------------
RMSE from ARIMA model is 33.97
By taking these parameters (4,1,2) and (4,1,2,12) SARIMA results are as under.
SARIMAX Results
====================================================================================
Date: Sun, 07 Mar 2021 AIC 581.322
Time: 14:22:21 BIC 609.983
- 132
==============================================================================
coef std err z P>|z| [0.025 0.975]
------------------------------------------------------------------------------
ar.L1 -0.9742 0.199 -4.899 0.000 -1.364 -0.584
ar.L2 -0.1122 0.285 -0.394 0.694 -0.670 0.446
ar.L3 -0.1044 0.277 -0.377 0.706 -0.647 0.438
ar.L4 -0.1285 0.162 -0.794 0.427 -0.446 0.189
ma.L1 0.1605 328.137 0.000 1.000 -642.976 643.297
ma.L2 -0.8395 275.462 -0.003 0.998 -540.734 539.055
ar.S.L12 -0.1441 0.364 -0.396 0.692 -0.858 0.569
ar.S.L24 -0.3597 0.227 -1.587 0.113 -0.804 0.085
ar.S.L36 -0.2153 0.106 -2.039 0.041 -0.422 -0.008
ar.S.L48 -0.1195 0.093 -1.281 0.200 -0.302 0.063
ma.S.L12 -0.5159 0.343 -1.503 0.133 -1.189 0.157
ma.S.L24 0.2086 0.373 0.559 0.576 -0.523 0.940
sigma2 215.3512 7.07e+04 0.003 0.998 -1.38e+05 1.39e+05
===================================================================================
Prob(Q): 0.86 Prob(JB): 0.30
===================================================================================
RMSE from SARIMA model is 17.54
ACF graph for Sparkling wine sales as under
PACF graph for Sparkling wine sales as under

It can be observed from above ACF and PACF plots, the cut off points for p and q for ARIMA model is 3 and 2 respectively.
By taking these parameters (3,1,2) ARIMA results are as under

ARIMA Model Results
==============================================================================
Dep. Variable: D.Sparkling No. Observations: 131
Date: Sun, 07 Mar 2021 AIC 2228.928
Time: 14:58:55 BIC 2249.054
Sample: 02-01-1980 HQIC 2237.106
- 12-01-1990
=====================================================================================
coef std err z P>|z| [0.025 0.975]
-------------------------------------------------------------------------------------
const 5.8816 nan nan nan nan nan
ar.L1.D.Sparkling -0.4422 nan nan nan nan nan
ar.L2.D.Sparkling 0.3075 7.77e-06 3.96e+04 0.000 0.308 0.308
ar.L3.D.Sparkling -0.2503 nan nan nan nan nan
ma.L1.D.Sparkling -0.0004 0.028 -0.013 0.990 -0.055 0.054
ma.L2.D.Sparkling -0.9996 0.028 -36.010 0.000 -1.054 -0.945
Roots
=============================================================================
-----------------------------------------------------------------------------
AR.1 -1.0000 -0.0000j 1.0000 -0.5000
AR.2 1.1145 -1.6595j 1.9990 -0.1559
AR.3 1.1145 +1.6595j 1.9990 0.1559
MA.1 1.0000 +0.0000j 1.0000 0.0000
MA.2 -1.0004 +0.0000j 1.0004 0.5000
-----------------------------------------------------------------------------
RMSE from ARIMA model is 1375.10
By taking these parameters (3,1,2) and (3,1,2,12) SARIMA results are as under.
SARIMAX Results
==================================================================================
Date: Sun, 07 Mar 2021 AIC 1219.260
Time: 15:01:53 BIC 1245.462
- 132
==============================================================================
coef std err z P>|z| [0.025 0.975]
------------------------------------------------------------------------------
ar.L1 -0.7556 0.151 -5.013 0.000 -1.051 -0.460
ar.L2 0.1169 0.185 0.633 0.527 -0.245 0.479
ar.L3 -0.0520 0.143 -0.365 0.715 -0.332 0.228
ma.L1 0.0330 0.191 0.173 0.863 -0.341 0.407
ma.L2 -0.9670 0.156 -6.197 0.000 -1.273 -0.661
ar.S.L12 -0.7538 0.496 -1.520 0.128 -1.725 0.218
ar.S.L24 -0.6371 0.351 -1.818 0.069 -1.324 0.050
ar.S.L36 -0.2469 0.151 -1.641 0.101 -0.542 0.048
ma.S.L12 0.3719 0.491 0.758 0.448 -0.590 1.334
ma.S.L24 0.3466 0.365 0.949 0.343 -0.370 1.063
sigma2 1.79e+05 1.67e-06 1.07e+11 0.000 1.79e+05 1.79e+05
===================================================================================
Prob(Q): 0.93 Prob(JB): 0.00
===================================================================================
RMSE from SARIMA model is 329.53
8. Build a table (create a data frame) with all the models built along with their corresponding parameters
and the respective RMSE values on the test data.
RMSE for Rose for Exponential Smoothing Models

Test RMSE
RMSE for Rose for ARIMA/SARIMA

Test RMSE
ARIMA(3, 1, 3) 15.99
SARIMA(0, 1, 2)x(2, 1, 2, 12) 16.52
ARIMA(4, 1, 2) 33.97
SARIMA(4, 1, 2)x(4, 1, 2, 12) 17.54
RMSE for Sparkling for Exponential Smoothing Models

Test RMSE
Test RMSE
RMSE for ARIMA Models for Sparkling

Test RMSE
ARIMA(2, 1, 2) 1375.03
SARIMA(0, 1, 2)x (0, 1, 2, 12) 321.48
ARIMA(3, 1, 2) 1375.10
SARIMA(3, 1, 2)x(3, 1, 2, 12) 329.53
9. Based on the model-building exercise, build the most optimum model(s) on the complete data and
predict 12 months into the future with appropriate confidence intervals/bands.
Since Rose data set has clear component of seasonality SARIM. Therefore, SARIMA model with parameters
(0,1,2)x(2,1,2,12) is selected for forecasting time line series and model details are as under.
SARIMAX Results
===================================================================================
Dep. Variable: Rose No. Observations: 187
Date: Sat, 06 Mar 2021 AIC 1191.208
Time: 13:41:48 BIC 1212.142
Sample: 01-01-1980 HQIC 1199.714
- 07-01-1995
==============================================================================
coef std err z P>|z| [0.025 0.975]
------------------------------------------------------------------------------
ma.L1 -0.8254 0.080 -10.334 0.000 -0.982 -0.669
ma.L2 -0.0807 0.086 -0.934 0.350 -0.250 0.089
ar.S.L12 0.0635 0.160 0.398 0.691 -0.249 0.376
ar.S.L24 -0.0340 0.019 -1.790 0.074 -0.071 0.003
ma.S.L12 -0.6953 0.207 -3.360 0.001 -1.101 -0.290
ma.S.L24 -0.0547 0.150 -0.365 0.715 -0.348 0.239
sigma2 166.0900 17.899 9.279 0.000 131.008 201.172
===================================================================================
Prob(Q): 0.79 Prob(JB): 0.02
Diagnostics of the Final Model for Rose
RMSE of full model is 33.47
The predicted sales of Rose for next 12 months is as below

Rose mean mean_se mean_ci_lower mean_ci_upper
1995-08-01 42.984338 12.890006 17.720391 68.248285
1995-09-01 43.513258 13.085191 17.866755 69.159761
1995-10-01 45.491994 13.141097 19.735918 71.248071
1995-11-01 57.520151 13.196775 31.654948 83.385355
1995-12-01 84.989586 13.252239 59.015674 110.963498
1996-01-01 20.575007 13.307260 -5.506743 46.656757

Rose mean mean_se mean_ci_lower mean_ci_upper
1996-02-01 30.224797 13.362216 4.035335 56.414260
1996-03-01 36.974058 13.416865 10.677486 63.270631
1996-04-01 38.520738 13.471317 12.117442 64.924035
1996-05-01 29.043623 13.525605 2.533925 55.553321
1996-06-01 36.323188 13.579682 9.707500 62.938875
1996-07-01 49.477037 13.633557 22.755757 76.198318
Since Sparkling sales data has component of seasonality. Therefore, SARIMA model with para meters (0,1,2) (0, 1, 2, 12)
is proposed to used for forecast for next 12 months using full data. Details of model are as under.
SARIMAX Results
===================================================================================
Dep. Variable: Sparkling No. Observations: 187
Date: Sun, 28 Feb 2021 AIC 2184.006
Time: 18:26:36 BIC 2198.958
Sample: 01-01-1980 HQIC 2190.081
- 07-01-1995
==============================================================================
coef std err z P>|z| [0.025 0.975]
------------------------------------------------------------------------------
ma.L1 -0.9094 0.104 -8.713 0.000 -1.114 -0.705
ma.L2 -0.1316 0.087 -1.507 0.132 -0.303 0.040
ma.S.L12 -0.5456 0.065 -8.393 0.000 -0.673 -0.418
ma.S.L24 -0.0202 0.084 -0.241 0.810 -0.185 0.145
sigma2 1.419e+05 1.32e+04 10.755 0.000 1.16e+05 1.68e+05
===================================================================================
Prob(Q): 0.91 Prob(JB): 0.00
===================================================================================
Diagnostics of the Final Model for Sparkling

RMSE of the Full Model is 550.06
The predicted sales of Sparkling for next 12 months is as below
Sparkling mean mean_se mean_ci_lower mean_ci_upper
1995-08-01 1874.535574 390.423218 1109.320128 2639.751020
1995-09-01 2487.730812 395.512564 1712.540432 3262.921192
1995-10-01 3299.133285 395.812900 2523.354256 4074.912313
1995-11-01 3937.427323 396.113009 3161.060093 4713.794554
1995-12-01 6136.305467 396.412891 5359.350478 6913.260456
1996-01-01 1251.541103 396.712549 473.998795 2029.083412
1996-02-01 1583.924557 397.012002 805.795332 2362.053782

Sparkling mean mean_se mean_ci_lower mean_ci_upper
1996-03-01 1842.202832 397.311309 1063.486976 2620.918689
1996-04-01 1822.837693 397.610284 1043.535857 2602.139529
1996-05-01 1668.252244 397.909036 888.364864 2448.139623
1996-06-01 1619.268043 398.207564 838.795558 2399.740528
1996-07-01 2021.196789 398.505870 1240.139637 2802.253941
10. Comment on the model thus built and report your findings and suggest the measures that the
company should be taking for future sales.
The decomposition of the time series trend of Rose wine is as below
Trend in sales of Rose is continuously decreasing over the period. Detailed study may be required to see whether
decreasing trend is due to change in customer preference or due to substitution. Seasonality of sales is observed, and
higher sales is maintained in the end of the year. Some promotion schemes and improvement / quality enhancers in the
product can be examined so as to attract new young generation customers.
The decomposition of the sales of Sparkling is as below.

Sales in Sparking does not have uniform trend but increased in some years and decreased later. Business study may be
done to find why sales are not increasing and what the contributing factors. Study can also include to see which wine
product has substituted/ had higher sales in the years of low sales of Sparkling. With promotion and focussed effort with
micro detailing it may be feasible to increase the sales. Sales of Sparkling wine higher in the later part of the year. This
may be due to climatic condition of the geography under study.

Rose Sparkling Wine

Uploaded by

Copyright:

Available Formats

Rose Sparkling Wine

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rose Sparkling Wine

Uploaded by

Copyright:

Available Formats

What trends are observed in the sales of Rose wine over time?

What trends are observed in the sales of Rose wine over time?

What observations can be made about the sales of Sparkling wine over time?

What observations can be made about the sales of Sparkling wine over time?

For this particular assignment, the data of different types of wine sales in the 20th century is to be

Rose wine sales

Sparkling -wine sale

Rose wine sales

Decomposition of Rose-wine sales into Trend, Seasonal and Residual

Distribution of sale of wine-Rose in each year

Trend for each month over the years in sale of wine-Rose

Distribution of sale of wine-Sparkling in each year

Distribution of sale of wine-Sparkling in each Month

Various models and forecast as below.

Rose wine sales

Sparkling wine sales

Results of Dickey-Fuller Test:

since p-value> 0.05, at alpha 0.05, time series is not stationary.

Results of Dickey-Fuller Test:

Results of Dickey-Fuller Test:

Rose wine sales

Lowest AIC is: 1273.194 with param (3,1,3)

ARIMA Model Results

Evaluation of model using RMSE: RMSE from ARIMA model 15.99

param seasonal AIC

107 (0, 1, 2) (2, 1, 2, 12) 774.969119

215 (1, 1, 2) (2, 1, 2, 12) 776.940114

323 (2, 1, 2) (2, 1, 2, 12) 776.996102

269 (2, 0, 2) (2, 1, 2, 12) 780.716942

161 (1, 0, 2) (2, 1, 2, 12) 780.992971

89 (0, 1, 1) (2, 1, 2, 12) 782.153872

322 (2, 1, 2) (2, 1, 1, 12) 783.703652

197 (1, 1, 1) (2, 1, 2, 12) 783.899095

95 (0, 1, 2) (0, 1, 2, 12) 784.014096

311 (2, 1, 2) (0, 1, 2, 12) 784.140949

Lowest AIC is : 774.969 with param (0,1,2) and seasonal (2,1,2,12)

Evaluation of model using RMSE : RMSE from SARIMA model 16.52

Sparkling wine sales

Lowest AIC is 2232.783 with param (2,1,2)

ARIMA Model Results

Evaluation of model using RMSE: RBSE for ARIMA model 1375.03

param seasonal AIC

95 (1, 1, 2) (0, 1, 2, 12) 1382.347780

41 (0, 1, 2) (0, 1, 2, 12) 1382.484254

101 (1, 1, 2) (1, 1, 2, 12) 1384.137874

149 (2, 1, 2) (0, 1, 2, 12) 1384.317618

47 (0, 1, 2) (1, 1, 2, 12) 1384.398867

107 (1, 1, 2) (2, 1, 2, 12) 1385.688721

53 (0, 1, 2) (2, 1, 2, 12) 1386.023734

155 (2, 1, 2) (1, 1, 2, 12) 1386.097242

161 (2, 1, 2) (2, 1, 2, 12) 1387.627785

77 (1, 1, 1) (0, 1, 2, 12) 1398.756167

Lowest AIC is : 1382.347780 with param (1,1,2) and seasonal 0,1,2,12)

Evaluation of model using RMSE: RBSE for ARIMA model 321.48

Rose wine sales

ACF graph for Rose wine sales as under

PACF graph for Rose wine sales as under

By taking these parameters (4,1,2) ARIMA results are as under

ARIMA Model Results