What and Why do Organizations Forecast?
—
Xiaorui(Jeremy) Zhu
08/25/2019
It is a prediction or estimate of an actual outcome expected in a future time period or for another situation.
The purpose: inform the process of planning.
The purpose of planning: a course of action so that your decision are based on more reliable forecasts.
Weather Forecast:
Stock Price:
Purpose:
Information:
Value:
Analysis:
System:
Evaluation:
Costs:
Sales/Activities
Social and Political Trends
Projects
—
Depends on the Problem Area
Forecasting object
Forecasting horizon
Forecasting frequency
Structure derived from decision making process
Forecasts required to support planning and decision making
Typically, we assume that the observations are equally spaced in time.
A set of comparable observations ordered in time. The values may refer to either a point in time (“the current value of the Dow Jones Index is 17,000”) or an aggregate over a period of time (“total sales for the last month were 350 units”).
Trend: A time series contains a trend if it shows systematic movements (e.g., increase or decrease) over an extended period.
Seasonality: A time series has a season component if it displays a recurrent pattern with a fixed and known duration (e.g., months of the year, days of the week).
Cycle: A time series has a cyclical component if it displays somewhat regular fluctuations about the trend but those fluctuations have a periodicity of variable and unknown duration, usually longer than one year (e.g., a business cycle).
Computer Software
Data Sources
The Federal Reserve Bank of St. Louis maintains an extensive database of U.S. and international macroeconomic series. The website is known as FRED
EconStats contains a large amount of economic and financial data on the United States and on several other countries, including the United Kingdom, China, and Japan.
Data.gov.uk offers easy-to-access government statistical series for the United Kingdom.
Econdata contains regional socioeconomic data for the United States. The Forecasting Principles website (http://www.forecastingprinciples.com/) provides a variety of links to other sources and data sets.
WRDS provides researchers with one location to access over 350 terabytes of data across multiple disciplines including Accounting, Banking, Economics, ESG, Finance, Healthcare, Insurance, Marketing, and Statistics.
The Nuts’n’Bolts hardware store stocks some high-value items and a large number of relatively low-value components. Inventory is reviewed weekly, and orders are placed with suppliers when inventory levels indicate that current levels are “too low.” Most suppliers deliver the low-value items within two weeks; some of the high-value items may take four weeks for delivery. How would you develop a forecasting system to predict the sales of such items at the individual SKU (Stock Keeping Unit) level? How would you use such forecasts to plan future purchases?
Use PIVASE to go through the necessary steps. For purposes of the discussion, assume that any historical data you might need can be made available.
The data file for U.S. retail sales (in Blackboard) which also contains monthly data for the Consumer Price Index (CPI).
Recreate Figures 1.2(time series plot) and 1.3(seasonality adjusted series) using R
Create a series called “real prices” by dividing sales by the CPI; then replicate Figures 1.2 and 1.3 for this series using JMP.
Compare the four graphs commenting on similarities and differences.