IS2000 | Principles of Information Science
  • Home
  • Schedule
  • Policies
  • Labs
    • Lab - SQL, SQLite
    • Lab - R
    • Lab - Decision Analysis
  • Resources
    • Lecture Notes >
      • Information Science as a Discipline
      • Information Modeling, Ontologies, UML
      • Information Architecture
      • Information Theory
      • Essential Excel
      • Essential R
      • Information Representation in XML
      • Patterns XML & DTD
      • Information Retrieval with XPath
      • Information Transformation with XSL
      • Information Storage in RDB
      • Information Retrieval with SQL
      • Information Retrieval Specification Algebras
      • Information Processing in R
      • Information-Based Decision
      • Information Management & Policy
      • Information Security & Privacy
      • Information Transmission
      • Information Systems
      • Predictive Analytics
      • Process Analytics
      • Information Visualization
    • Text Books
    • Tools
    • Additional Readings
    • Archive
  • Help
    • Blog

Practice Problems: Predictive Analytics


Exercises to practice predictive analytics, and construction and evaluation of forecasting models.

Akron Zoological Park

​During the late 1980s, the decline in Akron’s tire industry, inflation, and changes in governmental priorities almost resulted in the permanent closing of the Akron Children’s Zoo. Lagging attendance and a low level of memberships did not help matters. Faced with uncertain prospects of continuing, the city of Akron opted out of the zoo business. In response, the Akron Zoological Park was organized as a corporation to contract with the city to operate the zoo.

The Akron Zoological Park is an independent organization that manages the Akron Children’s Zoo for the city. To be successful, the zoo must maintain its image as a high-quality place for its visitors to spend their time. Its animal exhibits are clean and neat. The animals, birds, and reptiles look well cared for. As resources become available for construction and continuing operations, the zoo keeps adding new exhibits and activities. Efforts seem to be working, because attendance increased from 53,353 in 2003 to an all-time record of 133,762 in 2012.

Due to its northern climate, the zoo conducts its open season from mid-April until mid-October. It reopens for 1 week at Halloween and for the month of December. Zoo attendance depends largely on the weather. For example, attendance was down during the month of December 1995, which established many local records for the coldest temperature and the most snow. Variations in weather also affect crop yields and prices of fresh animal foods, thereby influencing the costs of animal maintenance.

In normal circumstances, the zoo may be able to achieve its target goal and attract an annual attendance equal to 40% of its community. Akron has not grown appreciably during the past decade. But the zoo became known as an innovative community resource, and as indicated in the table, annual paid attendance has doubled. Approximately 35% of all visitors are adults. Children accounted for one-half of the paid attendance. Group admissions remain a constant 15% of zoo attendance.

The zoo does not have an advertising budget. To gain exposure in its market, then, the zoo depends on public service announcements, the zoo’s public television series, and local press coverage of its activities and social happenings. Many of these activities are but a few years old. They are a strong reason that annual zoo attendance has increased. Although the zoo is a nonprofit organization, it must ensure that its sources of income equal or exceed its operating and physical plant costs. Its continued existence remains totally dependent on its ability to generate revenues and to reduce its expenses.

Construct a predictive model that provides an estimate for expected revenue for the next year.
Akron Zoo Data | Model Solution (WMA) | Model Solution (Trendline & 95% CI)

​Emergency Call Center

This attached Excel workbook contains call made to a 911 system for the past 24 weeks. For this data set, calculate forecasts for the next four weeks.
  • Construct a scatter plot. Do you see patterns or trends in the data? Which forecasting models would be most appropriate?
  • Build a Weighted Moving Average forecast using a four week span and adjustable weights. What weights would you choose? Why?
  • Build a time series forecasting model using simple linear regression. Are other regression models perhaps better.
  • Evaluate each forecast's fit using the MAD. Compare the models. Which provides better predictions?
  • Calculate the 95% CI for each forecast.​
Winter Park Emergency Call Center Data | Video Lecture Explaining Solution

RealSell Real Estate Sales

RealSell is a local real estate agency who wants to price client's offerings properly and competitively so they need a price forecast model based on various factors. The models should be a multiple regression model taking all relevant factors into account.
House Sales Data | Model Solution
Picture

Learning

Blackboard
Lynda.com

Support

Contact Instructor
FAQ
Terms of Use
© COPYRIGHT 2016-20 by Martin Schedlbauer
​FREE FOR ACADEMIC USE WITH ACKNOWLEDGEMENT. 
ALL RIGHTS RESERVED.