Method Selection

This section will help explain how Vesta users may choose and apply an analysis. We will use simple language to help explain the different applications for the available analysis methods in Vesta.

I want a way to explain the cause and effect of factors with relationship to an independent variable (such as number of cancer cases.)

If this statement explains your end-goal, refer to the Regression material.

My data behaves differently over a range of time frames. There is a variation in observed rates.

If you think there is a time-related relationship in your data that behaves differently over time (not a steady rate over time), refer to the Joinpoint material.

I want to estimate the values at locations that do not have observations.

This is a question of spatial interpolation and we recommend starting there.

I want to know whether my data forms clusters or contains outliers across my study area.

If you're looking for statistically significant hot spots, cold spots, or outliers relative to their neighbors, refer to the Local Cluster Analysis material.

I have many variables and want to see how they relate to each other, or simplify them into a smaller number of dimensions.

Refer to the Multidimensional Scaling material.

I think the relationship between my variables changes across my study area, not just its overall strength.

Refer to the Geographically Weighted Regression material.

I want to predict a yes/no outcome.

Refer to the Logistic Regression material in Regression, or the classification material in Machine Learning.

I want to predict an outcome from several variables, without assuming a specific mathematical form for how they relate.

Refer to the Machine Learning material.

I want to detect where and when unusually high (or low) rates occur.

Refer to the Space-Time Scan Statistic material.

I want to know whether my cases cluster in space and time together, not just one or the other.

Refer to the Space-Time Interaction Tests material.