Statistics

SpaceStat currently offers several ways to explore and analyze your data. You can explore the data in a more descriptive manner by taking the difference of two specified datasets, or the same dataset at two specified times, calculating custom datasets, or by standardizing data with a z-score transformation, or converting data into a different point or polygon geography. You can also smooth the data.

Several of the tools for data visualization, such as the histogram, scatter plot, and box plot, also provide descriptive statistics about your dataset (e.g., means, medians, standard deviations, etc.).  In addition, the graph statistics window for histograms provides values for coefficients of variation, skewness, and kurtosis, while the scatter plots graph statistics window presents values for simple and rank-based of correlation coefficients.  

You can examine population-level disparities in health outcomes using disparity statistics, or you can evaluate the spatial pattern through global and local clustering statistics.

• Global clustering statistics describe the study region as a whole.

• • Besag and Newell's statistic r

  • Moran's I.

• Local clustering statistics detect clustering in parts of the study area, with these there is a statistic for every location in the geography.

• • Besag and Newell's statistic l

  • Turnbull's statistic

  • Local Moran

  • Local G Test

  • Geary's C

Finally, you can use aspatial regression and geographically weighted regression tools to assess variation in one variable (the dependent variable, y) at set levels of another variable or variables (independent, or x variables).  Three forms of regression are included in SpaceStat:  traditional linear models, Poisson models, and logistic models.   You can also perform exploratory aspatial model building with model selection tools.