Transcript SSRG PowerPoint Presentation

Welcome!
• Who is Survivor SRG?
Brendan Biamon, Matthew Carroll,
Samira Fazel, Jesse Safir
• What is Survivor SRG?
Survivor SRG is a data mining and management organization that uses SAS
software and programming to provide statistical data of value.
In other words, we look for correlations between disparate data sets in an
attempt to map out previously undiscovered relationships.
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
Problems & Objectives
• Survivor SRG is looking at crimes in the U.S. and other currently
non-related data in an attempt to find meaningful correlations.
• Establishing such relationships can help in efforts to lower these
crimes and may offer insight into criminal behaviors.
• We chose ten different data sets to see if we could find any
correlations, whether they be significant or not.
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
The “Crime” Definitions
• The seven categories of crimes Survivor SRG used are those that the
U.S. Federal Bureau of Investigation tracks. They are:
• Assault
• Burglary
• Larceny
• Murder
• Rape
• Robbery
• Automobile Theft
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
Our Independent Variables
• We want to see if any of the following categories of information have
possible relationships to the seven categories of crimes:
• Divorce by actual numbers.
• Income per capita.
• Internet usage and purpose.
• Driver’s License totals by sex.
• Mental Health expenditure by total dollar amount.
• Population concentrations by square mileage.
• Annual Precipitation.
• Unemployment rates.
• Van registration.
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
The Data Sources
• Crime Data from the Federal Bureau of Investigations
• Divorce Data from the U.S. CDC
• Income Data from the U.S. Dept. of Commerce, Bureau of Economic Analysis
• Internet Usage Data from the U.S. Dept. of Commerce:
• License Data from U.S. Census
• Mental Health Data from the U.S. Department of Health and Human Services
SAMHSA
• Precipitation Data from the U.S. Dept. of Commerce, NOAA
• Square-mileage data from Enchanted Learning
• Unemployment Data from the U.S Bureau of Labor Statistics
• Van Ownership Data from U.S. Census
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
PROC Methodology
• Proc Chart
• Proc Gchart
• Proc Print
• Proc Plot
• Proc Corr (using EG)
• Proc Reg
• The REG procedure fits linear regression models by least-squares.
• Output: Proc Stepwise – forward; backward; stepwise
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
Rules & Assumptions of
Multiple Regression
• The independent variables can predict the dependent variable, but the
dependent variable cannot be used to predict the independent variables. It
is a one-way analysis.
•Independent variables should be justified theoretically.
•Independent variables selected should have strong correlations with the
dependent variable but only weak correlations with other independent
variables.
•Each independent variable has the same relationship with the dependent
variable at each value of other independent variables.
•Dependent variables' scores come from a normal distribution.
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
When do you use Multiple
Regression?
•Multiple regression is used when you want to determine what variables
contribute to the explanation of the dependent variable and to what degree.
•A common analysis is stepwise regression.
•In forward stepwise regression, independent variables are entered one at a
time to analyze how much each one adds to the explanation of the dependent
variable.
•In backward stepwise regression, independent variables are removed one at a
time and the effect on "R" (overall explanation of the dependent variable) is
evaluated.
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
Dependent Variable
Total Crime
Murders
Rape
Robbery
Aggravated Assault
Larceny
Motor Vehicle Theft
Independent Variables Best Suited for
Model
Percentage of Unemployed
Mental Health Spending per Capita
Percentage who Use the Internet
Mental Health Spending per Capita
Divorces per Capita
Percentage who Use the Internet
Percentage of Unemployed
Income per Capita
Mental Health Spending per Capita
Vans per Capita
Total Inches of Rainfall
Divorces per Capita
Income per Capita
Total Inches of Rainfall
Percentage who Use the Internet
Percentage of Unemployed
Total Inches of Rainfall
Percentage of Unemployed
Income per Capita
Mental Health Spending per Capita
Divorces per Capita
Total Inches of Rainfall
Vans per Capita
Survivor Statistical Research Group
“We get the job done, even if it kills us.”
Appendices
• Online Resource Library:
http://www.ibiblio.org/mcarroll/gschool/inls110-sas/appendices.html
• Our original data sources and SAS program sets
• Our SAS program code
• Our Project and results in total.
Survivor Statistical Research Group
“We get the job done, even if it kills us.”