You are on page 1of 5

OPIM-274 Business Forecasting

Fall 2022

Homework Assignment 1-Solutions

1)
a) There is need for direct forecast of variables such as number of consumers and their consumption habits, share of the market,

and any industry related trends (for example, increased or decreased interest in a “healthy” soft drink). You will also benefit from forecasts

related to the components (e.g., maintenance and utility costs), ingredients (e.g., price of fruits, sugar, and other ingredients), as well as

other related topics (e.g., packaging costs, marketing costs, etc.).


b) It is very likely that demand for these products is seasonal for two reasons: one is seasonality in consumption behavior (e.g., cold

drinks tend to have higher demand in summer time). Another reason is the seasonality of some of the retail outlets (e.g., sales of vending

machines are low during holidays, etc.).


c) Yes, it does as it dictates demand and the supply chain. Carbonated soft drinks or sodas are not perishable, and are generally safe

past the date stamped on the container, the quality and taste of the products may be affected by storage (perhaps, reduced taste or reduced

“fizziness”). Demand may be affected by the shelf life (stores putting the products that are near the end of their shelf life on sale or return

them to the distributor, if possible).


d) Mainly, we need data on sales of products (which may be complicated to get point of sales data due to multiple retail outlets) so

the order and distribution data is our best source for this. Also, consumption data is impossible to get for the orders that are made in bulk.

Alternatively, consumer surveys may be conducted to better understand consumption behavior. We also need to capture consumption

behavior and demand using marketing strategies. Finally, data on expenses and cost of services is also needed.
2) Answers vary
3) MSFT data

© Copyrighted Material-Professor Ali Arab


Interpretation: There’s an increasing, approximately linear trend over time (linear up until around mid-February 2020, followed by a declining trend for a

few weeks (which seem to correspond to the start of the pandemic), and again the trend switches to an overall increasing starting around April of 2020;

There is also some variability but no seasonality is apparent.


4) Problem 2.4

a and b)
The standard Excel output shows that the median is unchanged but the mean is much smaller and the standard deviation much larger. A few outliers can greatly

distort the analysis.

Returns   Returns with outlier  

Mean 7.48 Mean -3.08

Median 6.20 Median 6.20

Standard Deviation 9.96 Standard Deviation 37.39

MAD 7.60 MAD 21.70

c) These statistics only make sense for the investor if she has equal amounts invested in each stock.

5)
a) Uber rides:

Increasing trend, slightly nonlinear. Seasonality not clearly apparent.

Precipitation:

© Copyrighted Material-Professor Ali Arab


Mostly zero values with sporadic and seemingly random non-zero values. No specific trends or seasonality.

Tmax:

Increasing trend, slightly nonlinear. Seasonality not clearly apparent.

Tmin:

© Copyrighted Material-Professor Ali Arab


Increasing trend, slightly nonlinear. Seasonality not clearly apparent.

Wind:

Slightly decreasing, linear. Seasonality not clearly apparent. Higher variability in the earlier part of the data.

b)

Statistics

Variable Mean StDev Median

uber 78842 15844 78343

MAD: sum(abs(d_i)/n= 2269444.83/181= 12538.37

Mean is larger than the median, so we suspect the data may be slightly skewed to the right (due to potential large outlier(s)).

© Copyrighted Material-Professor Ali Arab


We may confirm this using a histogram:

© Copyrighted Material-Professor Ali Arab

You might also like