Lab 4: Final Project

Introduction: The research question I proposed for this final project was "Where should I buy a house in Outagamie County, Wisconsin?". This is the county that I grew up in and the data sets that I chose were based off of where I would want my house to be located if I was looking to start a family. The objectives for this project were to take the geoprocessing skills that I have learned over the course of this semester and apply them in several different ways as a means of answering my spatial question. There were certain parameters that had to be met such as 1) at least four tools had to be used (excluding clip, query, add field, project, and project define), 2) ideally, the area of interest (AOI) was to be at the county level, and 3) existing data and a minimum of three different spatial data layers (excluding the AOI) had to be used. We were also strongly encouraged to use the Mastering ArcGIS data, data provided by Esri, and/or data from the Wisconsin Department of Natural Resources (WDNR) which were readily available on the department server. I believe anyone who has just started a family or is looking to start a family and wants to live near or within the Appleton area in Outagamie County would be interested in this information, and it is these people who this information is intended for. 

Data Sources: The data and corresponding metadata I used to complete this project are as follows:

1. WiDNR2014.DBO.County_Boundaries: Wisconsin County Boundaries
2. Esri2013.DBO.blkgrp_usacen: U.S. Census Block Groups
3. Esri2013.DBO.mjr_hwys_usadata: U.S. Major Highways
4. Esri2013.DBO.rail100k_usadata: U.S. National Transportation Atlas Railroads
5. Esri2013.DBO.gschools_usadata: U.S. Geographic Names Information System Schools
6. Esri2013.DBO.park_dtl_usadata: U.S. Parks

The Esri data was provided with ArcGIS and the WDNR data came from the WDNR database at http://dnr.wi.gov/maps/gis/metadata.html.

There were a few data concerns that materialized as I worked through this project. One was with the completeness of the data set, in particular with the schools feature class. As I was looking through the public schools in the attribute table I noticed that Einstein Middle School, a large Middle School in Appleton, was not included in the data. This made me concerned that other public schools were not featured as well. Another concern I had was with the block group data set. I wanted to find the density of families based on the 'Families' and 'SQMI' fields as a way of finding neighborhoods with large percentages of families with children. However, it is hard to tell what the dynamics are of 'Families' with the U.S. Census data so I had to make an educated assumption that the likelihood of people who are reporting that their household is considered a "Family" is based on their being children in the household. 

Methods: There were numerous steps that I took to get the final result that I was looking for. I first dropped in the Wisconsin county boundaries data set and then selected and exported Outagamie County to get outagamie_co. This is the feature class I used to clip the rest of the data sets that I used in this project. Using the schools data I first created two features classes, cnty_grade_schools and cnty_middleandhigh_schools, set a four mile buffer on both, and then intersected the resulting classes so that I would have a proximity to both grade schools and middle and high schools. With the block group data set, I added a family density field which I then calculated using the 'Families' and 'SQMI' fields. I then selected and exported the data that produced high family density block groups (outagamie_highfamdens). Using the parks data set I set a one and a half mile buffer so that my final result would be within this distance of any park (outagamie_parks_final). I wanted my final result to be a certain distance away from both highways and railroads to cut down on the potential high level of noise, so I set a half mile buffer on major highways, a one mile buffer on railroads, and then combined the two results using a union (outagamie_hwyrail_buffer). Now that I had my base feature classes I desired it was time for the finalizing steps. I intersected the schools, high family density, and park feature classes to get areas that featured all three data sets (outagamie_parks_fam_schools), and then with that result I performed an erase with the highways and railroads feature class (potentail_house_area). To get a clean looking result, I then used the dissolve and clip tools which answered my question with potential house areas. The data flow model below summarizes my workflow through this project (Figure 1).

 

Figure 1: Data Flow Model

Results: The final result I achieved can be seen below (Figure 2), and I included an enlarged version of the county map for better clarity as well (Figure 3). Outagamie County is the red square-shaped figure in the locator map. The small, black symbols represent schools, the orange lines represent major highways, and the brown, hashed lines are the railroads. The tan polygons with gray outlines are the block groups, the green polygons are parks, and the purple polygons represent my final result of potential areas to live based on my criteria. The final resulting areas are all concentrated in the Appleton city area.

Figure 2: Final Result Map

Figure 3: Final Result County Map

Evaluation: I was overall pleased with how smoothly this project went, for the most part, and that I was able to find the data sets that I was looking for and could get the final result that I desired. The data provided by Esri and the WDNR was sufficient enough to allow me complete this project, although the data was not 100% accurate (the absence of Einstein Middle School for instance). If I had to do this project again I would possibly try to find a better method to figuring out which areas have a high density of families with young children such as using different parameters other than block groups and a different field aside from 'Families'. A faced a couple of minor challenges along the way. First, there were several different highway data sets to choose from so I had to be sure I chose the right one. I believe the major highways feature class was the right choice for the answer I was trying to attain. Another challenge was I had to use a trial and error method of figuring out my buffer distances. Too much buffer and I wouldn't have large enough areas in my answer; too little buffer and my areas would be too large. In the large scope of the project, these were minor challenges that I overcame without too much difficulty. 

Lab 3: Vector Analysis

Goal: The goals of this lab were to 1) learn how to use a number of geoprocessing tools for vector analysis and apply them to determine a suitable habitat for bears in an area of Marquette County, Michigan, 2) generate a data flow model, and 3) introduce a brief lesson on how to use python.

Background: Using GPS coordinates of bear locations within the study area of Marquette County, we must determine which forest types the bears are most commonly found in, how many of them are found near streams, the areas where the forest types intersect with Michigan DNR managed land, and how to eliminate areas that are near urban and built up lands.

Methods: The bear locations were saved in an excel file that contained its X and Y coordinates. Since these coordinates were not in a spatial database, they had to be added to ArcMap as an "event theme" which temporarily displays them as X,Y data. Once these locations were mapped, they had to be exported and brought into the geodatabase as a feature class. The next step was to determine what land cover type each bear location fell into in order to figure out which habitat type the bears were most commonly found in. A simple, inside spatial join brought together the bear locations and the land cover types tables. A summarize of the Minor_Type field (forest type) showed that most of the bears were located in either mixed forest land, forested wetlands, or evergreen forest land. Selecting these land types and creating a new layer from the selection produced the bear_landcover feature class (Figure 1).
The next step was to determine how many bears were found within 500 meters of streams. A 500 meter buffer of the streams produced the feature class streams_500m which was then dissolved to become streams_500m_dissolve. From this analysis, we found that 72% of bears were found within 500 meters of streams.
Based on the criteria we found up to this point (bear_landcover and streams_500m_dissolve), we now wanted to find suitable areas of bear habitat. This was done by performing an intersect between the two previously mentioned feature classes which then produced bear_streams_landcover. Since there were internal boundaries within this result, I used the dissolve tool to remove these boundaries which then resulted in the bear_streamlandcover_dissolve feature class.
Now that the suitable land was determined, I needed to figure out which sections of this land intersected with DNR managed lands. I added the dnr_mgmt shapefile to ArcMap and performed a clip with study_area to produce dnr_studyarea which showed the DNR managed lands located within the study area. Then I intersected dnr_studyarea with bear_streamlandcover_dissolve which resulted in the feature class dnr_bear_streamland.
The final objective was to exclude the areas that were within 5 kilometers of urban or built up land. My first step was to use select by attributes using landcover as the layer, Major_Type as the field, and "urban or built up areas" as the value. I then created a layer from this selection to give me the urban feature class. Next, I put a 5 kilometer buffer on this feature class to give me urban_5km. Finally, I used the erase tool to perform an erase between dnr_bear_streamland (input) and urban_5km (erase features) which resulted in the bear_habitat_final feature class.
After this was complete, I was introduced to python and briefly explored its functionalities. Using python, I performed a few function including buffer, intersect, and erase which can be seen below (Figure 2).
           

Figure 1: Data Flow Model

Figure 2: Python code

Results:

Figure 3: Map of Suitable Bear Habitat in Marquette County, MI

 The final suitable bear habitat areas on the upper map in Figure 3 appear in a golden color. As you can see, these areas are few and not very large in size. The cream colored sections (Land Cover/Stream) show the intersection of suitable land cover types and proximity to streams. The dark blue jagged lines represent the streams and the red dots are the locations of the bears. The outline of the upper map is the study area that appears in red on the lower location map where Marquette County appears in light green around the study area and both are located in the upper peninsula of Michigan.

Sources:

 (n.d.). Retrieved December 4, 2015, from http://gis.michigan.opendata.arcgis.com/  

Michigan 1992 NLCD Shapefile by County. (n.d.). Retrieved December 4, 2015, from http://www.mcgi.state.mi.us/mgdl/nlcd/metadata/nlcdshp.html 

Wildlife_mgmt_units. (n.d.). Retrieved December 4, 2015, from http://www.dnr.state.mi.us/spatialdatalibrary/metadata/wildlife_mgmt_units.htm 

Michigan Geographic Framework: Marquette County. (n.d.). Retrieved December 4, 2015, from http://www.mcgi.state.mi.us/mgdl/framework/metadata/Marquette.html