Visualizing Survey Data: Uncertainty, Outliers, and More

• Activity Number: 213
• Type: Invited
• Date/Time: Monday, August 5, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Government Statistics Section
• Author(s): Wendy L. Martinez*+ and Richard Holden
• Companies: Bureau of Labor Statistics and Bureau of Labor Statistics
• Keywords: Graphics ; Exploratory Data Analysis ; Survey Methodology ; Outlier Detection ; Non-response ; Spatial Data

Abstract

We will present several approaches for visualizing survey data. Some of these are traditional techniques being applied in novel ways, and others are new methods for dynamically exploring complex and high-dimensional data sets from surveys. For example, we will show how these have been used for graphical exploratory data analysis of survey non-response over space and time, the detection of outliers in high-dimensional data, the investigation of prediction intervals in time series, and more. These applications will be illustrated with data from the Office of Compensation and Working Conditions (OCWC) and the Office of Employment and Unemployment Statistics (OEUS).