a tool for performing meta-analyses and comparing results
PDFs represent parameter estimates for the US and the Bayesian aggregated result, of the effect of a standard deviation change in climate on violence, in terms of a z-score. Single parameter distributions: An unconditional probability density function, which may take any form, input either as a spline or a sampled function.
The relationship between log energy consumption and daily temperature from US data. 95\% confidence intervals are shown. Dose-Response parameter: A non-linear response, with a conditional PDF over the parameter values for every value of the "dose" variable.
Impact from various categories of hurricanes, and from tornadoes, floods, and severe storms. Error bars show 95\% confidence intervals. Categorical independent variable: When the independent variable takes discrete or categorical, the model is represented as a bar graph.
Portion of farm area by state, for cotton, maize, soybeans, and wheat. Only states with at least 25% of crop area allocated to those four crops are shown. Categorical dependent variable: When the probability distribution is defined over a discrete dimension, the model is represented as a probability mass function.
Bayesian estimate of log mortality as a function of temperature. The confidence intervals on μ are wide, reflecting the uncertainty in resolving the two estimates.
Relative yields for a given growing season average temperature.