I am looking at pulling data from NOAA flat files and API’s to try to build geo-located data sets which might be analyzed in an effort to build a predictive model for losses based on extreme weather events.
My main interest is in seeing the effects of extreme weather and understanding better how these events impact physical, financial and human ecology in a given location.
Initially I will draw data from these two separate repositories of data within National Oceanographic Atmospheric Association’s many data stores.
Because this is a preliminary analysis to see if building a predictor model based on weather data and reported storm losses is feasible, the initial model will include a limited area (such as a small state or parts of a larger state with more common severe weather occurrences).
I intend to complete a thorough preliminary analysis, build repositories to store the various data sets and shape files, create some kind of summary data to get a feel for the underlying structure of it and set up an initial model to try and predict losses based on weather data. Hopefully a dynamic map of the results or a plot of average weather conditions as appropriate and map probabilities of losses by region will help explain these preliminary model results.