Advanced GIS Lab9
Advanced GIS Lab9
Embedded Files
Objectives
Objectives
1) Determining whether data is dispersed or clustered
2) Determining the strength of clustered based on the associated value
3) Identifying statistically significant clusters of high and low values at each location
Introduction
Introduction
This spatial study assessed to find the significance of the spatial pattern and distribution of the data. Using the Arc GIS Geostatistical Analyst, such as Nearest Neighbor tool, Moran’s I test, and Hot Spot Analysis, this lab was conducted to identify patterns and statistically significant distribution of Bigfoot sightings. First, I did the Nearest Neighbor tool to determine that the data is dispersed or clustered. Second, Moran’s I test used to figure out the strength of clusters. In the last part, I exploited the Hot Spot Analysis for determining statistically significant clusters at each location. I conducted this study using ArcGIS Map 10.7.1., and the data from the Bigfoot Researchers Organization (BFRO).
Methods & Results
Methods & Results
1) Nearest Neighbor
In the first part, I used the Nearest Neighbor tool for seeking information about the spatial distribution and average distance. In this process, Euclidean distance, which is the calculating distance between two points, was conducted as a distance method.
After running the tool, I got observed mean distance (19064,3701 meters) and the expected mean distance (47325.4297). According to the result summary, the average nearest neighbor ratio is less than 1 (0.402836), and thus we can identify the result displays the clustering pattern. Also, this result is statistically significant because it has high z-score (-65.3479 and low p-value (p <.001). Therefore, we can say this is not randomly distributed.
2) Moran’s I test
In the second part, Global Moran’s I tool used to measure spatial autocorrelation of the location and a data value and identify it clustered together or dispersed in space. During the process, the fixed distance bands were selected as a conceptualization of spatial relationships, and Euclidean distance was used as a distance method. For figuring out the data, whether the pattern is clustered, dispersed, or random, threshold distance needs, and we selected 1,000km as a threshold distance.
The spatial autocorrelation report shows that the Global Moran’s index is 0.0035555, and this can indicate that it is row standardized, which means not skewed data. Also, the Moran’s Index is a positive value, and thus it means that the data is likely to cluster spatially.
The result displays that expected index (-0.000306) and variance (0.000) and this difference is statistically significant because it has high z-score (6.134660) and low p-value (p <.001). Therefore, this shows that the analyzed attribute is not randomly distributed.
3) Hot Spot Analysis
In the last part, the hot spot analysis was conducted for identifying statistically significant spatial clusters of high values (hot spots) and low values (cold spots). In this process, a Fixed distance band and Euclidean distance were used, and a threshold was selected different meters, such as 1,500,000, 500,000, or 100,000 meters. The distribution of cluster pattern was changed based on different threshold distance.
In this case, applying 100,000 meters tend to make reliable results. The analyzing result displays that red points and blue points show statistically significant because it has high z-score and low p-value. For instance, a hot spot with a 99% confidence level shows high z-score (5.671946) and a small p-value(p<.001). However, the red hot spot represents a reliable high value, and blue cold spot represents non-reliable low value. Thus, this result indicates that the data has a spatial clustering of high values.
Threshold_95.pdfLab9_95.pdfDiscussion
Discussion
In this lab, spatial statistics for determining the significance of the spatial patterning and distribution of the data were conducted. The analyzing data is interesting because of the statistically significant values of the Bigfoot sighting in North America. During the process, I could learn there are some limitations.
References
References
1) Bigfoot_pts.shp (BFRO.org)
2) NorthAmerica.shp(DIVA-GIS.org)
3) US_county_population (U.S.Census Bureau)
4) Lab 09 of WATS 6920 course, USU
Page updated
Google Sites
Report abuse