Statistics
In this lecture, we went over some of the important statistical terms in order to understand how to properly dig right into the data. Below are a few snapshots of what we covered.
Why Statistics?
- to explore and analyze the data set
- to make predictions, to summarize
- to look for relations among data. If confident, we can make predictions
What kind of data exists?
NOIR
- Nominal
- Ordinal
- Interval
- Ratio
How to summarize?
- Measures of central tendency (ex. mean, median, midpoint)
- Measures of dispersion (std. deviation, range)
Looking for relations?
- Visualizations: table, graphics
- Quantitative approaches: correlations does not assume causation
- regression – you need a hypothesis of what the dependent and independent variables are
Regression analysis
Looks at the relationship between dependent and independent variables, or response and causal variables. In my lab write-up, I go into more depth about regression analysis. But what is important to consider in geography is that positive spatial autocorrelation is the norm, so spatial declustering may be required to really understand what is happening.
Health Geography
Using GIS in health geography can be beneficial. The following are some various approaches health geography takes;
- a proactive approach rather than a disease mapping approach
- look at why different people cluster together
- socio-economic status has a strong relation to environmental justice
- exposure assessment
- proximity: ex. buffers, which way does the wind blow?
- interpolation: a model based on how something happens. Ex, West Nile, understanding how mosquitos live and the generation of a mosquito so you can map as a function of degree days vs the generation of mosquitos.
- Landuse regression spatial modeling: basically a more sophisticated interpolated surface.