The lab for Week 5 took us into the world of spatial statistics. The Esri course we completed introduced topics that focused on looking at spatial relationships within a dataset, identifying patterns, and making choices about which analysis tools to use based on trends in the data. Exploring the data is the first step.
Using a map of Europe showing weather monitoring stations, with temperature as the data to be analyzed, I looked at spatial relationships using Geostatistical Analyst tools. I found the mean and median values, noticing if they were located close to one another. By examining patterns in a histogram, a normal QQ plot, a Voronoi map, and a semivariogram cloud, I found outliers in the data. However, overall there was a normal distribution of values. The last part of the process before selecting an appropriate analysis technique is to look at spatial trends.
As a result of exploring the data in these ways, the most appropriate choice for analysis seemed to be geostatistical interpolation. This was a good fit because the data is normally distributed, stationary, autocorrelated, and has no local trends. If a weather station needed to predict future temperatures, this technique would be useful.
This lesson probably only showed the tip of the iceberg in terms of what ArcGIS is capable of, and it has gotten me thinking about how it could be used for a particular archaeology project I'd like to do in the future.
No comments:
Post a Comment