Statistical Analysis for Transformer Demand Forecasting

This document discusses the statistical tools and methods used for forecasting transformer demand and estimating the relationship between refrigerator sales and transformer demand. The appropriate tools for the analysis are inferential and descriptive statistical techniques, including hypothesis testing, correlation, regression analysis, and forecasting. The results are reliable and supported by various data points and analysis. The document also provides a due process for estimating transformer demand, which involves hypothesis testing, linear relationship estimation, and time series analysis. The regression analysis shows a good fit and presents a linear relationship between the given variables. The analysis would be helpful in estimation of demand, which can avoid the issue of understocking and overstocking.

STATISTICS
STUDENT ID:
[Pick the date]

ISSUE
Based on the given scenario, it is apparent that the there are two folds objective for
the operations manager Ratnaparkhi. One of these is to determine if there is a significant
change in the demand for transformers during the period 2009-2010. Additionally, a robust
forecasting model needs to be developed for the future forecasting of transformers required in
the future considering the sale of refrigerators.
STATISTICAL TOOLS & METHODS
A) The family of statistical tools that would be used for the analysis in the given scenario is
inferential statistical techniques. These are typically used when the objective is to test
various claims for the population based on the sample data. The appropriate tool used
would be dependent on the underlying objective and the distribution of the sample.
However, there is some use of descriptive statistics also in order to estimate the
relationship between variables along with conducting time series analysis. It is imperative
to note that descriptive statistical analysis tends to summarise the sample data (Eriksson &
Kovalainen, 2015).
B) The data provided in the given scenario is numerical in nature which has been represented
using the ratio measurement scale. With regards to quantitative variables, the descriptive
statistics techniques such as correlation and regression are useful. Also, the hypothesis test
can be performed in relation to the average and standard deviation pertaining to the
population.
C) The appropriate tool with regards to descriptive statistics would be correlation , regression
analysis and forecasting. With regards to inferential statistics, hypothesis testing is the tool
of choice.

D) Under the inferential statistics family, hypothesis testing has been deployed in order to
estimate if the demand for transformers has seen a change or not during 2006-2010.
Further, it has also been used for highlighting the significance of the linear regression
model estimated between refrigerator sales and transformers required. The demand
forecast of transformers in the future would become easier on account of regression
analysis as the future sales of refrigerators may be estimated and this can be used for
estimation of transformers. Besides, time series analysis of the transformer demand would
also be performed as it highlights the trend along with the seasonal component (Flick,
2015).
E) The regression analysis along with forecasting has been used to ensure that the estimation
of transformers is data driven and hence unlike the current practice of making random
guesses. This would enable reduction in overstocking and understocking with regards to
transformers and hence improve the overall inventory management. This estimation
technique would be an improvement over the current method where transformer
requirement is guessed using the previous two month estimates.
ANALYSE DATA & CONCLUDE
A) For estimation of the transformer demand, the due process would involve three steps. The
first step would be to test the claim if the average transformers required during 2006-2010
has changed or not. The second step is to estimate a linear relationship between
refrigerator and transformer demand. The final step is to perform a time series analysis to
forecast the future estimation of transformer demand.
B) The above due process is imperative as the hypothesis testing would help in understanding
if the average transformer demand has changed in the recent years or not. Further, the
linear relationship estimation would provide an estimate of transformer demand in the

future period based on the refrigerator sales projection. The time series analysis would
help in account for seasonal variations along with accounting for yearly trends.
C) The given results would be considered as reliable considering the fact that the estimation
of transformers demand is being supported through various data points and analysis. For
instance, the hypothesis testing and time series analysis both reflect changing average
annual transformer demand during the period from 2006-2010. Also, the seasonal aspect
has been accurately captured by time series analysis considering that data from multiple
years is available. Besides, the underlying relationship with refrigerators has also been
used to give precise estimates.
D) With regards to the change in the demand of transformers during 2006-2010, an ANOVA
analysis has already been performed using the transformer demand data from 2006-2008.
The relevant results are shown in Appendix 1. The relevant aspect is that the p value at
0.003 is lower than the assumed significance level of 1% which would lead to the
conclusion that the mean transformer requirement in atleast year differs from the others
(Fehr & Grossman, 2013). Considering the above conclusion drawn from 2006-2008 data,
it can safely be concluded that transformer demand does not remain the same across 2006-
2010.
In order to forecast the transformer demand in the future, the time series analysis is
found useful. Based on the transformer demand for 2006-2008, the time series graph has been
indicated in Appendix 2. It is evident that sales are on an upward trend on an yearly basis.
However, there is seasonality in the demand also which may be attributed to seasonal aspects.
It is apparent that the requirement of transformers tends to peak out in 6th and 7th month
(Hillier, 2016). These observations need to be considered for future estimation of
transformers demand.

The regression analysis has been conducted using sales of refrigerator as the
independent variable and transformer requirement as the dependent variable. The output from
Excel is pasted in Appendix 3. The regression equation is summarised below.
Transformer Requirement = 1239.84 + 0.31*Sales of Refrigerator
It is evident that the slope of the regression equation is 0.31 which implies that for a unit
increase in refrigerator sales, the requirement for transformer is expected to increase by 0.31
unit. Besides, the slope coefficient is found to be statistically significant which implies that
the underlying linear relationship between the given variables is significant (Hastie,
Tibshirani. & Friedman, 2014). Also, the coefficient of determination at 0.8548 is quite high
which highlights that 85.48% variation in transformer requirement can be estimated by
variation in sale of refrigerator. Hence, the given regression model presents a good fit (Hair,
Wolfinbarger, Money, Samouel & Page, 2015).
The above analysis would be helpful in estimation of demand which can avoid the issue of
understocking and overstocking. This would allow maximisation of profits and better
management of transformer related inventory.

References
Eriksson, P. & Kovalainen, A. (2015). Quantitative methods in business research (3rded.).
London: Sage Publications.
Fehr, F. H., & Grossman, G. (2013).An introduction to sets, probability and hypothesis
testing (3rded.). Ohio: Heath.
Flick, U. (2015). Introducing research methodology: A beginner's guide to doing a research
project (4thed.). New York: Sage Publications.
Hair, J. F., Wolfinbarger, M., Money, A. H., Samouel, P., & Page, M. J. (2015). Essentials of
business research methods (2nded.). New York: Routledge.
Hastie, T., Tibshirani, R. & Friedman, J. (2016). The Elements of Statistical Learning
(4thed.). New York: Springer Publications.
Hillier, F. (2016).Introduction to Operations Research.(6thed.). New York: McGraw Hill
Publications.

Appendix 1
Appendix 2
Appendix 3