Trends and Relationships to Explore
In regards to our previous blogposts, we will be analyzing trends of severe storms patterns based on geographical location, time and season, magnitude and economic damages. To narrow our analysis, we’ve decided to observe some trends and relationships of interest. Below are a few trends and relationships of interest we believe will be beneficial to our analysis:
- Pattern of severe storm based on sum of death toll
- West regions have a larger number of deaths and the northern part has a smaller amount of deaths.
- Pennsylvania, Texas, and New York are the three states with the most severe storms.
- Racial/socio-economic distribution of area suffered the most
- How climate change may increase or decrease severe storms
- Rising global temperatures is associated with widespread changes in weather patterns
- Investigating relationship between storms and car crashes in America
- Bad weather increases driving accidents
- Pattern between severe storms and plane delays in specific cities/regions
- Different storm types financial effect
- Which areas have what type of storms (midwest- tornados, west coast- earthquakes, east coast- hurricanes)
Initial Ideas for Modelling
To analyze these trends we’ve developed a few initial ideas for modelling. Some of our ideas include using the “plotly” function which produces graphs in Javascript and can be integrated with ggplot. We are also interested in potentially using an animated model with the “frame” function that is used to create animation. Furthermore, we were thinking about using an interactive table with DT or using the “crosstalk” function. We believe that creating interactive statistical models would be a good option for analyzing trends of weather patterns as it is a broad model. Using interactive approaches and different colors would be important when narrowing down a specific location, the type of weather pattern, and even determining a specific racial demographic.
Currently we have many ideas and do not have an initial fit of a model or results. We were thinking about potentially creating different models for different relationships of interests. In terms of what we need to do to proceed with our analysis, we will figure out how to convert the excel file into shp file. If we want to visualize with spatial map we would need to proceed with this research as the online converter is not working.