Skip to main content
Östersund map SaTScan Software for the spatial, temporal, and space-time scan statistics New York City map
Other Scan
Statistics Software
TreeScan, for the tree-based and tree-temporal scan statistics

SaTScan Tutorials

For educational purposes we have developed a set of SaTScan tutorial. These can be used in a class room setting, as part of a course on disease surveillance, spatial epidemiology, medical geography, or spatial statistics. They can also be used for self-study, or as part of a group project. With each tutorial comes a disease data set from New York State. The tutorials have been tested with a live audience as part of university courses and conference workshops. The only pre-requisites for these tutorials are a basic understanding of statistics and epidemiology at the undergraduate level.

SaTScan Tutorial #1: Purely Spatial Poisson Scan Statistic for Cancer Incidence

In this first SaTScan tutorial, the purely spatial scan statistic is used to analyze the geographical distribution of breast cancer incidence in New York State, in order to determine if there are any geographical clusters of breast cancer incidence. The tutorial teaches how to properly format and import the data, how to select parameter settings, how to run a purely spatial scan statistic with the Poisson probability model, how to interpret the results, and how to display the detected geographical clusters using Google Earth. The tutorial was designed for SaTScan v9.4, but can also be used with subsequent versions.

Authors:
Thomas Talbot, Sanjaya Kumar, Martin Kulldorff
Albany, NY, USA: November, 2014.
Download:
SaTScan Tutorial #1: Purely Spatial Poisson Scan Statistic for Cancer Incidence
New York State Cancer Incidence Data (45 MB)
Data Dictionary

SaTScan Tutorial #2: The Bernoulli Spatial Scan Statistic for Birth Defect Data

In the second SaTScan tutorial, the purely spatial scan statistic is used to analyze the geographical distribution of congenital malformation (birth defects) incidence in New York State, in order to determine if there are any geographical clusters of birth defects incidence. This tutorial walks you through the process of properly formatting and importing the data. It also shows you how to select parameter settings, how to run a purely spatial scan statistic with the Bernoulli probability model, how to interpret the results, and how to display the detected geographical clusters using Google Earth. The tutorial was designed for SaTScan v9.4, but can also be used with subsequent versions.

Authors:
Thomas Talbot, Sanjaya Kumar, Martin Kulldorff
Albany, NY, USA: November, 2015.
Download:
SaTScan Tutorial #2: The Bernoulli Spatial Scan Statistic for Birth Defect Data
New York State Birth Defect Data (1.5 MB)
Data Dictionary

SaTScan Tutorial #3: Advanced Options

In this third SaTScan tutorial, the purely spatial scan statistic is used to analyze the geographical distribution of female breast cancer incidence in New York State, in order to determine if there are any geographical clusters of breast cancer incidence. A purely spatial Poisson model will be used but we will describe and explore many of the advanced features available in SaTScan. While we illustrate these advanced features using the Poisson model, most of them are also available for the other probability models in SaTScan. The tutorial was designed for SaTScan v9.4, but can also be used with subsequent versions.

Authors:
Abdurrahman Abdurrob, Martin Kulldorff
Boston, MA, USA: September, 2016.
Download:
SaTScan Tutorial #3: Advanced Options
New York State Cancer Incidence Data (45 MB)
Data Dictionary

SaTScan Tutorial #4: Ordinal Scan Statistic for Identifying Unusual Cancer Stage Patterns

In this tutorial, we use the purely spatial scan statistic with the ordinal statistical model to analyze the geographical variation of colorectal cancer diagnosis in New York State, USA, in order to determine if there are any geographical clusters of either earlier or more advanced stage at diagnosis. That is, we will determine if there are any geographical areas where the distribution of cancer stage is unusually skewed relative to the statewide average. The tutorial was designed for SaTScan v9.4, but can also be used with subsequent versions.

Authors:
Francis Boscoe, Martin Kulldorff, Yueqing Wang
Albany, NY, USA: July, 2017.
Download:
SaTScan Tutorial #4: Ordinal Scan Statistic for Identifying Unusual Cancer Stage Patterns
New York State Colorectal Cancer Data (1.5 MB)

SaTScan Tutorial #5: Multinomial Scan Statistic for Identifying Unusual Population Age Structures

In this tutorial, we use the purely spatial scan statistic with the multinomial statistical model to analyze the geographical variation in age-specific populations in the United States in order to determine if there are any geographical clusters of populations which are unusually young, old, middle-aged, or any combination of these. The tutorial was designed for SaTScan v9.5, but can also be used with subsequent versions.

Authors:
Francis Boscoe, Martin Kulldorff
Albany, NY, USA: March, 2018.
Download:
SaTScan Tutorial #5: Multinomial Scan Statistic for Identifying Unusual Population Age Structures
2010 United States Population Data (< 1 MB)