Thursday, December 14, 2017

Food Deserts in Montgomery County

Time to wrap up our final project in Special Topics in GIS. This was kind of a combination of all the little things we learned in this phase; Food Deserts, Open Source software, Web map applications. For the final presentation I analyzed my own counties census tracts to identify food deserts, then presented my findings as both a web map application and a powerpoint presentation.



Tuesday, November 28, 2017

Food Deserts.. Again!

Similar to last week I'm doing food deserts again, but this time with my own data. I chose to analyze the county I live in, Montgomery County, TX. I obtained the counties census tract data from the Tiger/Line data website and used the same process as the previous week to isolate areas that could be food deserts. Using google earth I searched each portion of the county for grocery stores and used the location data from that to plot those points in GIS. With the grocery store points I could calculate the near distance to each census tracts centroid point, finding those that were farther than 1 mile from a grocery store. Surprisingly my county had a lot of food deserts. One of the problems I think is that Texas is very large and much of that area is undeveloped. It's entirely possible, especially in the north eastern part of the county, to drive for more than 10 miles without seeing an urban area. What surprised me the most is that the census tracts were large enough that even tracts covering of the major cities would still be considered food deserts just because of their size.

Like last week part of the lab was to use MapBox, Leaflet, and HTML to put my created layers on the internet for you all to see.

Online Map!

Tuesday, November 21, 2017

Food Deserts... On the web!

The next part of the food deserts lab sends me to the web, to learn how to use mapbox, leaflet, and HTML to put my map on the internet. There's not much to talk about this week, I learned a bit about HTML and how to use some free services on the web to share my map content. The web map I made shows the food deserts in Escambia county from the previous week classified by population density. The HTML code for the embedded map was copied from the page source of another map and the data was modified to point towards my map, and the legend, and geocoder were added with HTML. The map itself was created and hosted in mapbox.

Here's a link to the map.

Thursday, November 16, 2017

Food Deserts!

A food desert is an area where there is no large supermarket, one with fresh vegetables and meat, within easy travelling distance, usually one mile. Typically people in a food desert cannot afford to travel farther than one mile on a regular basis and back with groceries so their diet is usually deficient to a degree. The final project for special topics in GIS is fairly straightforward analysis but I'll be using an open source GIS program called QGIS to create the maps. QGIS is different than Arcmap in a lot of ways, but they perform the same basic functionalities. Creating a map layout seems to be a lot more difficult but other than that, the only differences seem to be the tools available and the layout of the program. My first task was to create a very basic map in QGIS to learn the program a bit.

A fairly basic map of Escambia county and the UWF campus. The second map I created used the census tract centroids in the study area and a CSV table of nearby grocery stores to create a map of food deserts and food oasis'.

The project will continue the analysis of food deserts in Escambia county and my actual project will be to do that analysis again myself on a county of my choosing.

Thursday, November 9, 2017

Photo Interpretation: Supervised Classification

Similar to last week, this weeks lab was about image classification. In this instance I used Supervised classification, which means the computer did a lot of the work. What it actually means is that I chose an area in the image that was identified and selected a portion of it to be classified. Then the software looked through the image and found other areas that it believed to be similar enough to the area I chose to be the same classification. The results of supervised classification are dependent on how many classifications you initially assign, the more data the software has the better your results. There are also tools for determining if areas of your image are likely to be classified incorrectly. My map this week includes a classified image as well as whats called the distance image. The distance image highlights pixels in the image that are farthest away from the classified pixels, meaning that they are potentially in the wrong class.

Friday, November 3, 2017

Photo Interpretation: Unsupervised Classification

Unsupervised Classification is the process of using a tool to group similar pixels in a raster into single classification to be later identified by an analyst. The tool examines a pixel and then compares it to the pixels around it. Depending on how different the surrounding pixel values are from the central pixel the tool will cluster those pixels together and classify them as a group. Once the tool has classified the image into the number of groups specified the analyst can then compare it to the original image and identify the classifications. For this lab the original image of the UWF campus in pensacola was classified in this way into fifty like categories. I then examined those categories and classified them into 5 general classes as practice.