Special Topics in GIS (GIS5935): Lab 12

In this week's lab, we delved more into how to utilize OSL analysis and GWR analysis on real life statistical data. Both analysis techniques allowed us to be able to visualize the areas that were more likely to call 911 based on the residents level of education and we are able to see if there was correlation in a particular crime based on certain variables, such as race and income.

OLS does not have any spatial component when performing a regression analysis unlike GWR which looks at the change over space the between the relationship of the the variables. I notice a major improvement in understanding the fluctuation of the prediction of calls to 911. The GWR was able to weigh the values in analysis to spatial location on the map.

Special Topics in GIS (GIS5935): Lab 11

In this week's lab, we got more in-depth with using regression analysis. The regression analysis can help determine our models by using the coefficient, P-value, Adjusted R-squared, AIC, and the VIF to see how the data correlates with our research question. If the VIF value is over 7.5, there is some redundancy in the data being used to correlate the desire result. Also, using the analysis of residuals can improve our model by seeing if the residuals are clustering in one areas. If residuals are cluttering, that means we need to add more variables to the analysis to get better results for the model.

Special Topics in GIS (GIS5935): Lab 10

How I got this is estimation is looking at the year 1950 through 2004 and notices that the rainfall tend to be less at Station A than Station B. I subtracted the value of the intercept coefficient of the regression analysis from the values of Station B. I believe that slope are going to increase as the years progress since climate change has a major factor in the amount of rainfall that these stations are receiving. The intercept points are probably going to very closer to each other since the values are not far off from each other.

Special Topics in GIS (GIS5935): Lab 8

IDW Interpolation

In this week's lab, we learned about different interpolation techniques and applied to water quality sampling points of Tampa Bay area. Three different interpolation techniques were used which were Thiessen, IDW, and Spline. The numerical results of both Thiessen and IDW were very similar; however the patterns of the results were quite difference from each. Thiessen retained it's polygon features when converting into a raster dataset unlike IDW which does not having any polygon features or outlines. The Spline interpolation technique shows the BOD concentration’s minimum is in the negatives and is higher than the previous interpolation techniques that were applied to this data. By having a higher minimum, the surface curvature minimum is that particular area to be extremely high since there are no sampling points in this area to create a smooth surface for the Spline interpolation. The Spline interpolation relies heavily on the input points to create a smooth surface. Lack of points in a particular location of the data will create a distorted when using Spline interpolation. To remedy this problem, deleting a point from the wq_bod data will reduce the minimum value of the points which will create a smooth surface.

 

Special Topics in GIS (GIS5935): Lab 7

In this week's lab, we explored TIN and DEM elevation models which we compare and contrast the models to each other. In the image above, shows the hard breakline of the bearlake.shp which is not present in the current data. This data was inserted by using the Edit TIN in ArcScene which the edges of the touching the lake feature replaced by hard edges. This allow for the outline shape of the lake to be present in the TIN data without needing the present of the lake shapefile dataset. The kind of technique is unique to TIN model since the lake feature's edge are imbedded into the model unlike the DEM model where the data would have to be on its own shapefile.