This document provides instructions for a time series forecasting assignment. It includes a table of monthly demand data from 2016-2020 and asks the student to:
1) Plot the time series data in two graphs and analyze trends, seasonality, and random variation.
2) Select a forecasting model and calculate a performance measure.
3) Use the model to forecast demand for 2021 and include a graph comparing forecasts to actuals.
4) Describe the selected forecasting approach and how it anticipates the analyzed behaviors, referring to graphs and performance measures.
5) Type answers to questions 1 and 4 and submit required files.
Part b (40 points)monthly time series forecasts starting jan. 202
1. PART B (40 points):Monthly Time Series Forecasts starting
Jan. 2021(Year 6)Due Date: Feb. 10th
1. Plot the monthly time series in Table 1 with the two types of
line graphs discussed in class. One graph should plot just one
line showing the change in demand across the years. The other
graph should plot one line for each year, showing the change in
monthly demand within the year.(All lines should be in same
graph for comparison for this second graph). Describe the
demand behaviors for this time series by referring to the graph
titles you labeled your two graphs with as you answer the
following questions:
A. Trend:
i. How would you describe the trend in this time series (e.g.
nonexistent, upward, downward, constant, growing faster/slower
than in past?):
There is no trend evident for the first four years, but at the start
of the year 2020 we can see a sharp upward trend in this time
series.
ii. Identify which graph(s) you looked at and explain how you
came to your conclusion based upon the graph(s):
The first graph is used to come to the conclusion above in
determining the trend because one single line needs to be seen
at a time to see the trend.
iii. Do you think the trend you described will continue next
year? Based on market signals, defend why or why not. What
type of trend behavior do you expect for next year? What other
information from the project description or news events today
did you consider in your analysis of this behavior?
It’s very difficult to say whether this uptrend will persist
because of lot of other factors like vaccination, rate of
vaccination, government lenience and its policy.
B. Seasonality:
i. How would you describe the seasonality in this time series
2. (e.g. not present, present with high and low periods in which
months?):
Seasonality is present for the first four years and can be seen
evidently from the graph 2. For the year 2020 there is a sharp
increase due to COVID-19.
ii. Identify which graph(s) you looked at and explain how you
came to your conclusion based upon the graph(s):
As mentioned earlier, graph 2 is used to understand the
seasonality.
C. Random Movement:
i. How would you describe the amount of random variation in
this time series (e.g. minimal, extremely volatile?):
There is seasonality seen for the first four years, but after that
there is an abrupt increase in the demand.
ii. Identify which graph(s) you looked at and explain how you
came to your conclusion based upon the graph(s):
Both the graphs are used to come to above mentioned
conclusions.
2. Identify a time series forecasting model that you think will
provide Mr. MMM with a reasonably accurate forecast for the
next year. Calculate at least one monthly measure of
performance (e.g. MAD or MAPD) to describe how accurate
your selected forecasting approach is.
Looking at the two graphs we can see that moving averages
would be appropriate for forecasting. We will calculate two
forecasts one 3 month moving average and 4 month moving
average and then calculate the measures of performance MAD,
MSE and MAPE (MAPD) and then decide on these factors
which method of forecasting needs to be used.
Based on the measures of performances mentioned, 3 month
moving average seems to be a better method of forecast due to
lesser error values.
3. Another method of forecasting that can be used is seasonality
index. The measure of performance used for this is coefficient
of correlation.
After calculation we can see that the value of coefficient of
correlation is 0.69, hence one can say this is a good forecast.
3. Using your recommended forecasting approach, forecast the
monthly demand for each month in 2021(year 6). Include a
third graph in your spreadsheet that depicts how your forecasts
(historical or future) compare to the actual patterns observed in
the time series. Be sure to title your graph with a meaningful
label that refers to the periods forecasted. The graph legends for
the lines should clearly distinguish between historical and
forecasted values.
The forecast for the next month is calculated for both 3 month
moving average, 4 month moving average.
The forecast for the next year is calculated in the excel sheet
along with the third graph.
4. What forecasting approach did you select to model this time
series? If you selected a seasonal method, describe clearly what
other time-series models you used to implement this approach.
Describe how your forecast for next year anticipates the
behaviors you described in part 1. You should refer to specific
relevant graphs and interpret the information the graph(s)
provides as well as your calculated measure of performance
when describing the quality of your forecast. Interpret your
measure of performance in a managerial context and be sure to
include the units assumed for your measures (e.g. %, $, time
period). You should use the results in a sentence that describes
how you would communicate the potential
uncertainty/inaccuracy to Mr. MMM as he considers using your
forecast. (You will not get full credit for just reporting the
numbers!)
4. We have used 3 month, 4 month moving average and seasonality
index method of forecasting. After implementing the seasonality
index we have done linear regression method to find the trend
and then seasonalize it using the earlier calculated the
seasonality index. The error measures are calculated earlier for
all the forecasting methods.
5. Type your answers to the questions in step 1 and step 4.
Submit your word document and your spreadsheet with the
forecast model you recommended in steps 2 thru 4 in
Brightspace’s Project 1 Part B Submission folder. Do not
include other models besides the one you are recommending.
Make sure your spreadsheet includes the three requested graphs.
Data20162017201820192020# Operating Weeks per
MonthJanuary22.4829.1632.5342.9534.424February22.0531.923
7.439.6740.914March32.7429.0833.438.7155.95April21.1624.9
720.9222.6850.354May20.421.1521.8722.9944.085June23.0624.
4420.3121.0641.484July20.1121.1522.4723.6646.944August24.
9520.6324.8621.8151.74September22.9821.1422.5922.5760.485
October30.7834.6134.0234.0265.524November32.2734.0233.55
34.5170.244December29.8325.631.013490.524
1. Graph20162017201820192020# Operating Weeks per
MonthJanuary22.4829.1632.5342.9534.424February22.0531.923
7.439.6740.914March32.7429.0833.438.7155.95April21.1624.9
720.9222.6850.354May20.421.1521.8722.9944.085June23.0624.
4420.3121.0641.484July20.1121.1522.4723.6646.944August24.
9520.6324.8621.8151.74September22.9821.1422.5922.5760.485
October30.7834.6134.0234.0265.524November32.2734.0233.55
34.5170.244December29.8325.631.013490.524N95 masks
ordered (in million)January, 201622.48February,
201622.05March, 201632.74April, 201621.16May,
201620.4June, 201623.06July, 201620.11August,
201624.95September, 201622.98October, 201630.78November,
6. 29.16 31.92 29.08 24.97 21.15 24.44
21.15 20.63 21.14 34.61
34.020000000000003 25.6 32.53 37.4 33.4 20.92
21.87 20.309999999999999 22.47 24.86
22.59 34.020000000000003 33.549999999999997
31.01 42.95 39.67 38.71 22.68 22.99
21.06 23.66 21.81 22.57
34.020000000000003 34.51 34 34.42
40.909999999999997 55.9 50.35 44.08 41.48
46.94 51.7 60.48 65.52 70.239999999999995
90.52
Month, Year
N95 masks ordered (in million)
N95 masks ordered (in million)
2016 January February March April May June July
August September October November December
22.48 22.05 32.74 21.16
20.399999999999999 23.06 20.11 24.95
22.98 30.78 32.270000000000003 29.83
2017 January February March April May June
July August September October November
December 29.16 31.92 29.08 24.97 21.15
24.44 21.15 20.63 21.14 34.61
34.020000000000003 25.6 2018 January February
March April May June July August September
October November December 32.53 37.4 33.4 20.92
21.87 20.309999999999999 22.47 24.86
22.59 34.020000000000003 33.549999999999997
31.01 2019 January February March April
May June July August September October
November December 42.95 39.67 38.71 22.68
22.99 21.06 23.66 21.81 22.57
34.020000000000003 34.51 34 2020 January
February March April May June July August
September October November December 34.42
40.909999999999997 55.9 50.35 44.08 41.48
11. 2. Go to the "Insert" option in the toolbar, select the "Line"
graph option and select the type of line chart you want to plot.
3. Right-click on the chart that appears on the sheet and click on
the "Select Data" option.
4. Click on the "Chart Data Range" box of the dialog box that
appears.
5. Select the data required and you will observe some lines
appearing in the graph. Click OK to confirm the same.
The chart plotted for the above data is as below:
Now we need to calculate the seasonal index.
The seasonal index is calculated as follows:
1. Calculate the mean (sum up all the values for that particular
month and divide by the number of years) of the data for each
month.
2. Now using the average values obtained from the data,
calculate the grand mean of these monthly means.
3. To find the seasonal index for each of the month divides the
individual mean of that particular month by the grand mean.
The calculations of the step by step procedure explained above
is as follows:
12. Now to calculate the forecast for required months, we need to
deseasonalize the demand by multiplying the seasonal index of
that particular month with the demand in that particular month
of any year. Note that for all the years any particular month has
the same seasonal index. The values for the same is as shown
below:
After deseasonalizing the demand we need to find the trendline
equation. And this is done as follows:
Starting from the first year number the periods as 1,2,3,4,5.... so
on and so forth. This is our independent variable (X) and the
deseasonalized demand is dependent variable (Y)
Solving the equations below the value of a and b required for
the trend line is as calculated. The equations are:
∑Y = n*a + b*∑X
∑XY = a*∑X + b*∑X^2
Trend line or regression line = Yc = a*x + b
where, Yc = Forecasted value
13. The other required values i.e ∑Y, ∑X, ∑(X*Y), *∑X, ∑(X*X)
are calculated and as are as shown below:
The value of a and b after solving the above-mentioned
equations and trendline equation is as follows:
The trendline equations is as follows:
Now for the fourth year forecast, we need to substitute the
various values of x starting from 37, 38, 39.... 48 and find the
corresponding demand. This is done because the periods for the
fourth year will continue as previously. Once the forecasted
demand using the trend line equation is found, we need to
multiply the corresponding month's (37 is for Jan month of
fourth year, 38 for Feb and so on) forecasted value with the
corresponding seasonal index. The calculations of the same are
as follows:
17. 27.416031787389272 28.305054301262018
24.7226452323718 28.026244282483042
26.877115613340511 24.092184853368735
43.577029940984687 37.812427352912678
33.42203278371877 26.536366362078816
28.875949753493241 26.480237054085155
28.867971165537615 24.832390760680795
24.700757100249294 28.026244282483042 27.
646177639832519 26.415165592213384
34.92249989682638 38.994363574682566
48.263798310769289 58.911201337331057
55.365457378598613 52.155756552870479
57.272297823767353 58.864493458376764
66.189711538461552 53.976470469967332
56.269704938331962 70.326493806092799
Month
N95 demand of mask in million
Graph 3 - Seasonalized demand, deseasonalized demand and
linear regression
Data (3)20162017201820192020# Operating Weeks per
MonthJanuary22.4829.1632.5342.9534.424February22.0531.923
7.439.6740.914March32.7429.0833.438.7155.95April21.1624.9
720.9222.6850.354May20.421.1521.8722.9944.085June23.0624.
4420.3121.0641.484July20.1121.1522.4723.6646.944August24.
9520.6324.8621.8151.74September22.9821.1422.5922.5760.485
October30.7834.6134.0234.0265.524November32.2734.0233.55
34.5170.244December29.8325.631.013490.524
DataNumber of U.S. coronavirus (COVID-19) cases from Jan.
20, 2020-Jan. 24, 2021, by dayNumber of new cases of
coronavirus (COVID-19) in the United States from January 20,
2020 to January 24, 2021, by day*
1/20/2051/21/2001/22/2001/23/ 2001/24/2011/25/2001/26/2031/
27/2001/28/2001/29/2001/30/2011/31/2012/1/2012/2/2002/3/20
32/4/2002/5/2002/6/2012/7/2002/8/2002/9/2002/10/2002/11/201
2/12/2002/13/2022/14/2002/15/2002/16/2002/17/2002/18/2002/
20. 1/21248,0891/12/21220,5281/13/21198,7881/14/21217,1661/15/
21225,5731/16/21226,6081/17/21246,4851/18/21212,2531/19/2
1185,3831/20/21142,2401/21/21152,9371/22/21187,9191/23/21
188,1761/24/21190,994
1 (a) Number of U.S. coronavirus (COVID-19) cases from Jan.
20, 2020-Jan. 24, 2021, by dayNumber of new cases of
coronavirus (COVID-19) in the United States from January 20,
2020 to January 24, 2021, by day*
1/20/2051/21/2001/22/2001/23/2001/24/2011/25/2001/26/2031/
27/2001/28/2001/29/2001/30/2011/ 31/2012/1/2012/2/2002/3/20
32/4/2002/5/2002/6/2012/7/2002/8/2002/9/2002/10/2002/11/201
2/12/2002/13/2022/14/2002/15/2002/16/2002/17/2002/18/2002/
19/2002/20/2002/21/20192/22/2012/23/2002/24/20182/25/2002/
26/2062/27/2002/28/2032/29/2003/1/2003/2/2023/3/20443/4/20
213/5/2003/6/20843/7/2003/8/2003/9/202593/10/202243/11/200
3/12/202913/13/202773/14/204143/15/20363/16/2003/17/201,82
23/18/203,5513/19/203,3513/20/204,7773/21/2003/22/2003/23/2
016,3543/24/2020,3413/25/2003/26/2016,4203/27/2003/28/2016
,8943/29/2018,0933/30/2019,3323/31/2017,9874/1/2022,5594/2/
2024,1034/3/2026,2984/4/2028,1034/5/2032,1054/6/2033,5104/
7/2026,4934/8/2029,5104/9/2031,7094/10/2030,8594/11/2035,3
864/12/2031,6064/13/2031,6334/14/2029,3084/15/2024,4464/16
/2025,8024/17/2028,7114/18/2032,5494/19/203 0,0234/20/2028,
2524/21/2027,6684/22/2025,6344/23/2024,0194/24/2029,1274/2
5/2030,7194/26/2038,5094/27/2032,4174/28/2029,2184/29/2022
,5414/30/2020,5175/1/2031,3795/2/2031,7745/3/2026,7535/4/20
31,8395/5/2029,2665/6/2016,2005/7/2022,2675/8/2022,1195/9/2
030,2045/10/2025,8705/11/2026,6425/12/2023,7675/13/2018,04
45/14/2021,4245/15/2020,8405/16/2027,0905/17/2022,8135/18/
2031,9675/19/2013,2275/20/2024,4175/21/2023,3105/22/2022,7
875/23/2020,4755/24/2024,1515/25/2026,1585/26/2015,2535/27
/2024,8865/28/2016,3625/29/2019 ,6065/30/2021,2145/31/2017,
9626/1/2023,4826/2/2026,1166/3/2014,6926/4/2024,8906/5/201
4,5836/6/2020,0696/7/2028,9226/8/2028,9186/9/2017,8486/10/2
017,5366/11/2017,2356/12/2020,3156/13/2021,7456/14/2022,13
36/15/2025,3146/16/2021,7546/17/2018,5146/18/2027,9216/19/
25. 51700000 60480000 65519999.999999993 70240000
90520000
Number of cases of COVID 19
Masks ordered
3Month valueNumber of casesMasks ordered in 2020 (in
mn)2020January11134420000February25540910000March31405
7455900000April486333450350000May573006644080000June6
80359641480000July7185093046940000August8151093851700
000September9117751160480000October10177571565520000N
ovember11423014770240000December12626391390520000SU
MMARY OUTPUTRegression StatisticsMultiple
R0.8851255227R Square0.783447191Adjusted R
Square0.7617919101Standard
Error7578786.37528132Observations12ANOVAdfSSMSFSignifi
cance
FRegression12.07799853744517E+152.07799853744517E+1536
.17811260360.0001295248Residual10574380029221497574380
02922149.7Total112.65237856666667E+15CoefficientsStandard
Errort StatP-valueLower 95%Upper 95%Lower 95.0%Upper
95.0%Intercept42476677.31578052949887.2590853414.399423
97290.000000051735903918.933035249049435.6985257359039
18.933035249049435.6985257X Variable
17.38209657571.22731705546.01482440340.00012952484.6474
63772610.11672937894.647463772610.1167293789Hence the
linear regression model is y = 42476677.32 + 7.38*xFor January
2021 the value of x is 13, hence the forecast y = 42476677.32 +
7.38*13 = 42476773.26
PART A (20 points):Monthly Regression Forecasts starting Jan.
2021 (YEAR 6)DUE DATE: Feb. 3rd
1. Open the Coronavirus2020Data.xlsx spreadsheet posted on
the website. You will need to clean the data by performing the
following steps:
a. Make sure that the number of cases in column C are numeric
so that they can be properly added and used in your regression
analysis.
26. b. Use the Text to Columns menu option to divide the
month/day/year for each date into separate columns. Use the
Sumif formula on this modified data to convert the daily
demand into monthly demand estimates for 2020 so that you
have 12 data points for your regression analysis.
2. Plot a scatter plot of 3M’s U.S. N95 Monthly Mask Demand
(see Table 1 above)versus # Monthly Cases of Covid in U.S. in
2020. Make sure your graph and axes are properly labeled.
3. Run a simple linear regression of 3M’s U.S. N95 mask
demand on the # Cases of Covid in U.S. in 2020. In a textbox on
your spreadsheet, describe the relationship between 3M’s U.S.
mask demand and the # of cases in the U.S. Write the linear
regression model you would use to predict demand each month
next year for 3M. Experts are predicting that there will be
6,517,535 Covid casesthis January 2021. What is your forecast
for 3M N95 mask orders in January if that is the case?
Note: 1 to 4 of the Part (a) has been answered separately in the
excel spread sheet.
4. Do you think this is a good model to predict each month’s
demand for 2021? Defend your answer with numbers from your
regression analysis here:
Yes, this model is good in predicting the demand for the future
months, because as we can see in the linear regression analysis
provided the value of adjusted r is 0.88, which indicates that the
model is good enough with moderately little variability.
5. What is the difficulty of using this model to predict 3M’s
demand each month next year? Type 1-2 sentences to answer
this question here:
The difficulty that we might face in predicting the demand for
next year is when the demand doesn’t grow as per expectations
or in other words the demand may diminish exponentially or
may surge exponentially because of the case being dealt here.
27. 6. What other variable do you think could be used to explain the
monthly demand for 3M’s masks next year? Defend why you
think this variable should be considered and what type of data
you would need to collect to measure it. Type your answers
here:
The other variables that can be used are:
Number of people using the mask more than once in a day.
Number of people using the mask only once in a day.
Number of people reusing the mask.
7. Submit your word document with the typed answers to
questions 4-6 and your spreadsheet with your work for
questions 1-3 in Brightspace’s Project 1 Part A Submission
folder.
BCOR 3750 - Dr. Leon
January 26, 2021
PROJECT 1: 3M’s N95 Mask 2021 Aggregate Production Plan
3M is the leading U.S. producer of the N95 respirator masks
that are in high demand as a personal protective equipment
(PPE) item for U.S. healthcare workers and other at-risk
employees/individuals during the current pandemic. N95 masks
are considered the gold standard in personal protective
equipment because they block 95% of large and small particles
utilizing a unique electrostatic filter. Recently as new variants
of the coronavirus emerge, health organizations are encouraging
ALL citizens to replace cloth and surgical masks with N95 or
KN95 masks (see last page of project).
28. 3M’s PPE products for the U.S. market are currently produced
at its Aberdeen plant in South Dakota as one of its main product
lines for its health care business segment, which historically has
contributed about 25% of the U.S. net revenues for the
company. Mr. MMM, a hypothetical production manager at
3M’s Aberdeen Plant uses aggregate forecasts to assist in its
production planning for the aggregate mask product line. Table
1 shows 3M’s monthly U.S. mask orders for the last five years
along with the number of operational weeks anticipated per
month (last column) that it operates the plant.
Given recent surges in U.S. demand due to the pandemic and
criticism by government agencies for not providing the critical
supply, 3M invested $80 million in upgrades to its Aberdeen
production facility, increasing equipment capacity for this
product line two different times last year. Since October the
weekly plant capacity has been expanded to 22 million units.
Mr. MMM would like to forecast the aggregate monthly mask
product line demands for this upcoming year so that he can
adjust scheduled productions and workforce requirements
appropriately and make any necessary subcontracting
arrangements. He has asked you to analyze the historical data
to identify recent patterns that 3M’s healthcare segment should
take into account when it develops the N95 product lines’
production plan for the 2021 year (i.e. year 6).
3M currently employs 600 workers to run the upgraded
automated assembly line 5 days per week, 8 hours per day, who
are paid $14.50 per hour. The material and other variable costs
are $0.40 per mask on average. Every worker scheduled to
support the production line allows the plant to produce an
incremental 750 masks per hour. They can hire additional
workers to run the expanded line if they decide they want to
fully utilize the upgraded equipment’s capacity. Since 3M’s
workforce is unionized, 3M cannot resize its workforce freely
as new employees will become part of the union. It costs 3M
29. $5,000 to hire and train a new worker, and $10,000 to lay an
employee off given the union restrictions. If 3M asks their
employees to work overtime, the extra hours are scheduled on
weekends or evenings to get additional capacity time if
necessary. Employees are paid time and a half when they work
overtime.
3M currently does not have any inventory of its masks. The
holding cost for each mask is $0.05 per month. In the past, 3M
has just rejected mask orders if they could not meet the current
monthly demand. The company started backordering demand
this year as needed to help the U.S. during the pandemic. It
costs $0.10 per mask per month to process backorder requests.
At the end of December, 3M did not have any backorders in the
U.S.
Historically, Mr. MMM’s aggregate planning strategy has been
to maintain a steady, well-trained workforce throughout the year
in order to control costs, ensure quality production and keep the
union happy. The workforce works full-time and produces as
many masks as possible each month. Starting December, 3M
decided overtime labor may be considered as an option to
address the societal impact of mask shortages if their capacity is
fully utilized during the workday. Government regulations limit
employee overtime to no more than 20% of the regular
production level in any particular month.
Mr. MMM also has a longstanding relationship with a Chinese
manufacturer that subcontracts work from American companies.
The Vietnamese manufacturer insists that 3M subcontract at
least five million masks every month next year in the contract;
in return, this subcontractor will manufacture 3M’s masks for
$0.70 per unit, which includes the cost of the material, labor
and shipping.
Mr. MMM is unclear as to what type of demand he should
30. expect for the masks next year. He believes that the pandemic
is likely to continue for the near future, but the impact on
demand could greatly decrease once the vaccine is successfully
deployed to most of the population. He recognizes he may have
to make adjustments to his usual production strategy in order to
control costs with the recent expansions and improve product
availability simultaneously. The union will not allow him to
schedule the workforce to work a shorter week though (less than
5 days per week). He can resize the workforce and/or use the
subcontractor. While he is willing to use backordering now to
help the U.S. economy, he does not want to see a plan that has
any unfilled orders at the end of December 2021, when demand
can surge with the cold weather and other viruses. He also does
not want to have an inventory larger than 50 million masks at
the end of December 2021 in case the pandemic is under control
and future demand decreases as a result.
2016
2017
2018
2019
2020
# Operating Weeks per Month
January
22.48
29.16
32.53
42.95
34.42
4
February
22.05
31.92
37.40
39.67
33. Table 1: Historical number of 3M N95 masks ordered in the
U.S. over last 5 years (in millions)
In order to help 3M identify a good aggregate plan (Part C), it is
first necessary to identify a good forecast (Part A and Part B).
(Ignore the # of Operating Weeks per Month data in Table 1 for
Forecasts in Part A & B).
PART A (20 points): Monthly Regression Forecasts starting
Jan. 2021 (YEAR 6) DUE DATE: Feb. 3rd
1. Open the Coronavirus2020Data.xlsx spreadsheet posted on
the website. You will need to clean the data by performing the
following steps:
a. Make sure that the number of cases in column C are numeric
so that they can be properly added and used in your regressi on
analysis.
b. Use the Text to Columns menu option to divide the
month/day/year for each date into separate columns. Use the
Sumif formula on this modified data to convert the daily cases
into monthly case estimates for 2020 so that you have 12 data
points for your regression analysis.
2. Plot a scatterplot of 3M’s U.S. N95 Monthly Mask Demand
(see Table 1 above) versus # Monthly Cases of Covid in U.S. in
2020. Make sure your graph and axes are properly labeled.
3. Run a simple linear regression of 3M’s U.S. N95 mask
demand on the # Cases of Covid in U.S. in 2020. In a textbox on
your spreadsheet, describe the relationship between 3M’s U.S.
mask demand and the # of cases in the U.S. Write the linear
regression model you would use to predict demand each month
next year for 3M. Experts are predicting that there will be
6,517,535 Covid cases this January 2021. What is your forecast
for 3M N95 mask orders in January if that is the case?
4. Do you think this is a good model to predict each month’s
demand for 2021? Defend your answer with numbers from your
regression analysis here:
34. 5. What is the difficulty of using this model to predict 3M’s
demand each month next year? Type 1-2 sentences to answer
this question here:
6. What other variable do you think could be used to explain the
monthly demand for 3M’s masks next year? Defend why you
think this variable should be considered and what type of data
you would need to collect to measure it. Type your answers
here:
7. Submit your word document with the typed answers to
questions 4-6 and your spreadsheet with your work for
questions 1-3 in Brightspace’s Project 1 Part A Submission
folder.
PART B (40 points): Monthly Time Series Forecasts starting
Jan. 2021 (YEAR 6) DUE DATE: Feb. 10th
1. Plot the monthly time series in Table 1 with the two types of
line graphs discussed in class. One graph should plot just one
line showing the change in demand across the years. The other
graph should plot one line for each year, showing the change in
monthly demand within the year. (All lines should be in same
graph for comparison for this second graph). Describe the
demand behaviors for this time series by referring to the graph
titles you labeled your two graphs with as you answer the
following questions:
A. Trend:
i. How would you describe the trend in this time series (e.g.
nonexistent, upward, downward, constant, growing faster/slower
than in past?):
ii. Identify which graph(s) you looked at and explain how you
came to your conclusion based upon the graph(s):
iii. Do you think the trend you described will continue next
year? Based on market signals, defend why or why not. What
type of trend behavior do you expect for next year? What other
information from the project description or news events today
did you consider in your analysis of this behavior?
35. B. Seasonality:
i. How would you describe the seasonality in this time series
(e.g. not present, present with high and low periods in which
months?):
ii. Identify which graph(s) you looked at and explain how you
came to your conclusion based upon the graph(s):
C. Random Movement:
i. How would you describe the amount of random variation in
this time series (e.g. minimal, extremely volatile?):
ii. Identify which graph(s) you looked at and explain how you
came to your conclusion based upon the graph(s):
2. Identify a time series forecasting model that you think will
provide Mr. MMM with a reasonably accurate forecast for the
next year. Calculate at least one monthly measure of
performance (e.g. MAD or MAPD) to describe how accurate
your selected forecasting approach is.
3. Using your recommended forecasting approach, forecast the
monthly demand for each month in 2021 (year 6). Include a
third graph in your spreadsheet that depicts how your forecasts
(historical or future) compare to the actual patterns observed in
the time series. Be sure to title your graph with a meaningful
label that refers to the periods forecasted. The graph legends for
the lines should clearly distinguish between historical and
forecasted values.
4. What forecasting approach did you select to model this time
series? If you selected a seasonal method, describe clearly what
other time-series models you used to implement this approach.
Describe how your forecast for next year anticipates the
behaviors you described in part 1. You should refer to specific
relevant graphs and interpret the information the graph(s)
provides as well as your calculated measure of performance
when describing the quality of your forecast. Interpret your
measure of performance in a managerial context and be sure to
include the units assumed for your measures (e.g. %, $, time
period). You should use the results in a sentence that describes
36. how you would communicate the potential
uncertainty/inaccuracy to Mr. MMM as he considers using your
forecast. (You will not get full credit for just reporting the
numbers!)
5. Type your answers to the questions in step 1 and step 4.
Submit your word document and your spreadsheet with the
forecast model you recommended in steps 2 thru 4 in
Brightspace’s Project 1 Part B Submission folder. Do not
include other models besides the one you are recommending.
Make sure your spreadsheet includes the three requested
graphs.PART C Background Work: 3M’s Aggregate Plan Draft
DUE DATE: February. 17th
Complete Reading Assignment 1. You will not be turning the
reading assignment questions in, but they will help you think
about important aspects of the different tools you might choose
to utilize in your aggregate plan. In particular, reflect carefully
about the employees 3M will need to hire (e.g. skills they need
to have and the union impact), the available labor pool and
3M’s relationship with the subcontractor. Use this information,
your forecast from part B and the Aggrega te Planning
spreadsheet template posted in the Brightspace’s Project 1 Part
C module to develop an aggregate production plan for 3M by
performing the following tasks:
1. Analyze the cost and performance associated with the Mr.
MMM’s currently proposed aggregate plan at 3M. In this plan,
3M does not sign the contract with the subcontractor but keeps
all current 600 employees working 40 hours per week
throughout the year. He will not ask employees to work
overtime but instead backlog all unfilled orders. Enter this
strategy on the Status Quo worksheet of the provided Project1
Aggregate Planning template. Note potential problems that
could occur if Mr. MMM implements this plan next year with
respect to the goals he has in mind.
2. Still using your forecast from part B and the spreadsheet
model posted in the Assignment section for Project 1 Part C,
identify a new aggregate plan on the Recommended Plan
37. worksheet that you think would be a better strategy for 3M.
You may utilize any combination of the capacity tools
suggested by Mr. MMM in your plan for this project
assignment, including the subcontractor, resizing, backordering
and overtime, but you must abide by the stated conditions for
the tools. As you experiment with various combination of
capacity tools, consider not only the impact that the tools have
on the short-term total cost and constraints, but also on the
long-term competitiveness of the company and the company’s
objectives. You should refer to your aggregate planning
reading assignment and your assessment of the labor
requirements for possible long-term impacts. Pick one
aggregate plan to recommend to Mr. MMM. Note that it does
not have to be the plan with the lowest short-term cost. You
can justify your plan based on its overall long term and short
term attractiveness.
3. Create the following visualizations to describe important
information when you make your recommendation to Mr. MMM.
a. A table that identifies the changes you are proposing the
company makes for the different aggregate tools (e.g workforce
size, OT hours/week, units subcontracted, units backlogged).
You should be comparing the recommended usage from the
manager’s suggestion to your suggested plan in a table that does
not include unnecessary information.
b. One or more graphs that compare the performance of a key
performance indicator (e.g. costs, units backlogged, inventory
levels) for the two different strategies that highlights the
BENEFITS of your recommended strategy.
c. Draft a paragraph that describes why you think your plan is
better for 3M than the manager’s Status Quo plan and how it
will address problems/issues facing 3M and the U.S. this
upcoming year. Considering 3M’s product, industry and the
economy outlook for next year, what are the strengths of your
plan? How will your plan make 3M more competitive in the
long run? Refer to your tables and graphs to support your
38. points.
4. Submit your spreadsheet file with both aggregate plans (the
status quo and your recommended plan) and your part 3
visualizations into Brightspace’s submission folder for Project 1
Part C Aggregate Planby Wednesday, February 17th as material
for our Synchronous Project 1 Class Activity.
PART C (40 points): Final Aggregate Plan Recommendation &
Memo to Mr. MMMDUE DATE: Feb 24th
Following principles learned in the workshop, prepare a one
pagesingle-spacedtyped memo to Mr. MMM (Audience) that
provides a recommendation for the aggregate plan you think 3M
should follow next year. (Use 11 pt or larger font, 1 inch
margins and submit in Brightspace Project 1 Part C submission
folder).
Assume that Mr. MMM has NOT seen the aggregate planning
spreadsheet model you submitted which is very overwhelming
and hard to process. You should provide sufficient detail of
your plan in your memo so that Mr. MMM can implement your
recommendation without having to refer to your spreadsheet
model . To describe your plan, you may attach as many tables
and graphs as you like as additional pages to the memo. These
tables and graphs should be properly documented, referenced
properly in the memo and easy to read.
Your memo should be persuasive, concise and strongly
supported so that it convinces Mr. MMM that your aggregate
plan recommendation is better for 3M than the status quo plan
he is currently considering. To do this, you should make sure
you summarize/highlight the issues up front that he will
encounter with his status quo plan. You should analyze the
effectiveness of the selected tools implemented in your
recommendation by referring to the noted improvement in key
performance metrics (KPIs). You will also want to refer to your
reading assignment 1, anticipated forecast behaviors and the
company's policies regarding its workers and stock levels as you
39. organize the critical points you will use to "sell" your
recommendation. The points you decide to include in your
memo should describe the recommended final plan, NOT a
description of your thought process for how you came to this
conclusion. Use of tables and graphs examples presented in the
synchronous class activity would be helpful in illustrating
differences between the plans.
You have permission to change your aggregate plan
recommendation submitted before the Project 1 Class Activity
workshop based on feedback you received that day if you think
there are serious problems with the aggregate plan you
identified for that day. You must resubmit your updated
spreadsheet into the Project 1 Part C assignment link by
Monday February 22nd. The last aggregate plan submitted by
end of that day will be the aggregate plan that will be graded.
It isimportant that the memo you submit on the 24th describes
the last aggregate plan you submitted into Brightspace. Points
will be lost if these two pieces are not in sync.