Sunday, November 23, 2014

GIS4930 Module 4 - Tiling & Basemap Creation

Compilation of the grocery store layer included referencing several online sources plus verifying the information in Google Streetview. Online sources included Yellow Pages, Manta, and Google Maps which turned out not to be very reliable. Two grocery stores were shown in completely wrong locations. This was verified by personal knowledge and double-checking street view. Verifying grocery stores in the outlying rural areas involved more research to verify whether they were full-service or convenience stores (street view, store websites or descriptions, and reviews were used). Grocery stores outside the study area were included, because of the 10-mile marker used to define rural food deserts. (http://www.ers.usda.gov/data-products/food-access-research-atlas/documentation.aspx) Addresses for the grocery stores were tabulated in Excel, saved as csv file, and geocoded before performing the Near analysis to determine which urban census tract centroids exceeded the 1-mile marker and which rural census tract centroids exceeded the 10-mile distance from a grocery store.

Centroids of the census tracts were calculated to provide a comparison point for the one-mile marker food desert determinant (urban) and 10-mile marker (rural). Interestingly, the centroid for one of the rural census tracts landed inside one of the city census tracts. Both city census tracts were considered to be food deserts which surprised me. If one of the grocery stores had been located on the east end of town or farther south, at least one of the city census tracts probably would not have been classified as a food desert. The grocery stores are both located near the western edge of town (and extreme NW part of the larger urban census tract). Because of the town size and number of grocery stores, this type of analysis might not be as informative as anticipated. Perhaps an evaluation by census block would be more appropriate to really define the at-risk areas, or an assessment by residents’ age and mobility/transportation issues could be done. I was surprised that at least one of the two southern rural census tracts was not a food desert. However, the grocery stores in that area are of a smaller nature and do not carry stock nearly as comprehensive as a chain grocery store would. So while not technically food deserts, these rural areas might still be prone to a lower-quality diet unless supplemented with home-grown produce.

The process of creating a web map of food deserts included tiling the shapefiles described above. Tiling a shapefile enables more rapid loading of layers into a web map because the layers are displayed in smaller portions (tiles). In TileMill each of the data layers for this map was centered in a tight-fitting bounding box with a centralized marker. The zoom level was adjusted so that the resulting tiles would be 10MB. Each of these map layers had up to 19 tiles. The default world layer was removed by editing the code, and the symbology (colors, opacity) for the layers was adjusted as needed. Each layer was exported for use in Mapbox and by default was sent off to a folder not of my choosing. I opted to re-organize those layers into project folders prior to adding them to the map in Mapbox. The method of saving layers or projects in both TileMill and Mapbox was not very intuitive for me (save/don't save, save again, export, etc.), so several times some steps had to be repeated in order to ensure that I had the desired product. As so often happens, this was just enough exposure to new software to let me know how much there is to learn! The screenshot below shows the food deserts, census tract boundaries, and grocery stores added to the revised basemap in Mapbox.


Food Deserts in Marshall, Michigan
with Census Tract Boundaries and Grocery Stores
Created with TileMill and Mapbox

Sunday, November 16, 2014

GIS4930 Module 4 - Tilemill & Leaflet

The exploration of opensource software for GIS purposes continued this week with the focus on Tilemill and Leaflet. Shapefiles were imported into Tilemill which already had a basemap in place. Arcmap was used to classify the food deserts by population (QGIS could have been used also) before selecting the color scheme in Color Brewer. This color scheme was entered into the Tilemill source code. The Leaflet code editing was terribly confusing, but with a lot of referencing previous instructions, double-checking coding changes, verifying paths, and re-reading of instructions, I made it through. OpenCage Geocoder was utilized to find latitude and longitude for specific addresses. Markers and shapes with popup text boxes were edited in the code. Layer groups were created and the ability to toggle layers on or off was coded. geocoder was added to the source code for finding locations. Interestingly enough, it went to locations within a few blocks of the input addresses, but not to the exact address location specified. Its reliability indicates that it should be investigated further. Throughout the process of editing the Leaflet code, the text and html files were saved incrementally, keeping previous copies in case something really got messed up during the editing process. The gods of coding must have been with me this week though, because I made it through all the Tilemill and Leaflet coding without any errors. What a lovely surprise that was!

Here's the link to the final web map: Food Deserts of Southeast Escambia County which opens up this:
Food Deserts of Southeast Escambia County





Thursday, November 13, 2014

GIS4035 Module 10: Supervised Classification

ERDAS Imagine was used to develop two different land use maps during this module, with the method of supervised classification being the focus of the exercises. In areas of interest (AOIs), polygons were drawn around representative pixels of different land uses or land covers, or a polygon was "grown" from a seed pixel to encompass a sample of a land use or land cover area. These samples ("signatures") were then used to classify the remainder of the area covered by the image. The classes were critiqued using histograms and Mean (statistically speaking) Plots, and additional signatures were gathered as needed. (That was the intent anyway. Initially I thought the classification was looking fairly decent after some adjustments, but then I noticed something fishy about the roads, including the acreage attributed to roads. Obviously there is some other feature lumped in with roads which needs to be evaluated further.) After the additional signatures were incorporated, the histograms and Mean Plots were used to critique the classification again. Bands 3, 4, and 5 had the least amount of overlap, and so various combinations of them were used to edit the signature colors before finally settling on R3, G4, B5 for the Germantown map shown below. The image was classified with this band combination, and a distance file was created. The purpose of the distance file was to highlight specific areas which might have been misclassified. Similar classes were merged by the process of recoding, and areas were calculated.
Supervised Land Use Classification
Germantown, Maryland

Tuesday, November 11, 2014

GIS4930 Module 4 - Intro to QGIS (Prepare Week)

Prepare week for this module started off with an introduction to QGIS, an open source, desktop GIS. The similarities between QGIS and ArcGIS are significant, but there are several differences as well. QGIS did reveal a few quirks, one being that an added layer did not show up in its correct location, but what could possibly be bad about UWF being in the Bahamas at this time of year? UWF's location was easily corrected by removing and re-adding the layer to the map. A difficulty I encountered was not being able to find adequate help resources (in a timely manner) for issues I experienced while using QGIS. Adding and adjusting bar scales was much more involved in QGIS than it is in ArcGIS. The use of data frames in the composer was a particular challenge for me. Frequently my data frame would be void of information after changes were made elsewhere in the composer. Hopefully more experience will alleviate those issues.

All that being said, a definite advantage to QGIS is that it's free and available to anyone. This would be beneficial for small businesses and non-profits which do not have the financial resources available to invest in ArcGIS.
University of West Florida
(Introductory QGIS Map)
For this module's project, QGIS was used to perform several processes which have been utilized in ArcGIS. These included re-projection, clipping, selecting by attributes, selection by polygon (this turned out to be a challenge), and centroid generation. From this information, food deserts were calculated based on whether the centroids of the census blocks were within 1 mile of a grocery store.

Statistics were also generated in QGIS after using an edit session to delete unnecessary columns in the attribute table for the food desert shapefile and again for the non-food desert shapefile. The results of these processes are shown in the following map of Food Deserts and Non-Food Desert areas in Pensacola, Florida.
Food Deserts and Non-Food Deserts of Pensacola
Escambia County, Florida

Tuesday, November 4, 2014

GIS4035 Module 9: Unsupervised Classification

Unsupervised classification of land cover allows the software to group pixels according to similarity (in this case creating 50 different classes). While this simplifies the classification process, the software is unable to distinguish different surfaces which reflect similarly. Because of this the results of unsupervised classification need to be reviewed and adjusted by an analyst. For instance, some grassy athletic fields (identified by soccer nets and light fixtures) reflected similarly to paved surfaces and were classified accordingly during the process of unsupervised classification. After review of the classes created with unsupervised classification, the pixels grouped like this were manually assigned to a more appropriate class. The reviewing process was made easier with the use of the Swipe, Flicker, and Blend tools along with the use of color-highlighting of pixels. 100% accuracy is not expected with classification. Some pixel groups did not clearly belong to one particular class; these pixel groups were assigned to a Mixed class. A few dormitories appear to have sod roofs, but this is due to those pixels being classified as grass like the overwhelming majority of pixels in a particular class.

After the image was reviewed and classes manually adjusted, the classes were merged to reduce the number to five using Recode. In the Recode Table, an area column was created from which percentages of permeable and impermeable surface areas were calculated. In this case, the information was transferred to Excel for computations. The resultant map shows the recoded image along with acreages for each of the classes of land cover and the percentages of permeable and impermeable surface area.
Unsupervised Classification of Land Cover and Surface Area
University of West Florida Campus