Geoprocessing and Geospatial Analysis
Through coursework at the Centre of Geographic Sciences I have learned a lot about Chloropleth Mapping and Grouping Analysis including the pitfalls associated with not representing data with statistically appropriate groupings.
I am familiar with analyzing data geospatially as a large part of term one completion included learning and applying that tools and statistics associated with regression analysis including exploratory regression, ordinary least squares regression and geographically weighted regression. To properly analyze the data a thorough understanding of residuals, model fit and collinearity was required.
This is a study on the potential market for an ethnic grocery store in Halifax NS. It is not a full evaluation as it only considers 3 variables. Each variable was broken into 5 groupings and assigned a weight. The variable weight values were added up for each sample zone and a weighted suitability was calculated for primary, secondary and tertiary polygons.
The purpose of this portfolio piece was to consider 6 variables from European Countries that are present on the World Bank Data webpage and group them using different methods including : No spatial constraints, continuity edges corners, delauney triangulation, k nearest neighbours and spatial weights.
This was an exercise in stepping through the different stages of regressions in order to determine which datasets were statistically the best to model a chosen variable (in this case food). This process included exploratory regression, OLS regression and geographically weighted regression. The results from each step was used in the next. Model fit, residuals and collinearity were just some of the variables considered during the course of this project.
The goal of this project was to use the concentric rings method to analyze the trade area around a selected business (in this case ethnic grocery store). Proportional Sums and Averages were calculated to determine demographics and income etc in the first concentric ring.