This week combined what we learned in prepare week about Food
Deserts and QGIS and Analyze Week 1 where we utilized Mapbox and Leaflet.
The challenge this week was to select our own city and combine the skills
we learned in all of the software components to produce a final map in leaflet
viewable via a link directed to our I:/ drive. Here is the link to my
leaflet map: http://students.uwf.edu/rh51/GIS4930_SpecialTopics/leaflet/leaflet_rh.html
Here is an image
of the map but I recommend using the leaflet link to access the geocoder and
zoom functions:
The tiling process
essentially takes a continuous image and breaks it into portions known as
tiles. Each tile is about 256x256 pixels and since they are placed sided
by side they appear to be a seamless map to the web user. By
combining tiles and zoom levels the process is very effective for using in web
mapping applications. Tiled maps cache efficiently so that your web
browser can quickly recreate them from the cache instead of having to download
each time you want to view the map. Because they progressively load from
the outside to the inside you can go ahead and zoom in to a general area and
the map will catch up as the loading process is not interrupted. Tiles
are also simple to use and formatted for numerous technologies such as servers,
web, desktops and even mobile devices.
As part of our lab
we provided answers to some excellent "food for thought" questions.
I have included those here for parties interested in more detail of the
map making process.
1. Explain your data including the
source, processing required (did you create it? How?), and information about
the quality/credibility of your food deserts and grocery stores.
Initially I
downloaded as much data as I could think to get. I prefer to start with more, evaluate
the data and then choose what best fits the project. I used a combination of data from the
census.gov tiger site as well as FGDL. Because
I decided to use the Basic style provided in Mapbox I did not need to utilize
the road files I had previously downloaded, nor any information regarding
bodies of water. The data
that became most helpful was the census block data (2010), City limit polygon
for the City of Port Saint Lucie, county boundary polygons and grocery store
locations from google earth exported to a KML file. Once I determined the data I intended
to use, I selected a coordinate system that was available both in ArcDesktop
and QGIS as QGIS did not contain the initial Fl State Plane East I wanted to
use. I settled upon the UTM
Zone 17N (EPSG:26917) which was the best choice given my options in QGIS. Once this determination was made I
performed a batch project on my shp files in ArcCatalog and placed them in a
UTM17N folder. This became
my working folder from which I used QGIS to create my census block clips to
eliminate the riparian boundary portions of the properties. I converted the KML to shp files in
ArcMap and added my other files within UTM17N to review my data and confirm it was
all in the correct projection. I
also made sure to set the projection in QGIS. Within QGIS there was other data which
needed to be created from existing datasets. I clipped the Census Blocks for the
entire state to the City of Port St. Lucie shp file and defined the results as
my study area. I then
created Polygon Centroids for my Study area. In ArcMap I created a near file using
my imported Polygon Centroids by making them the input Feature and my Grocery
Stores as the near feature. Since
the City of Port St. Lucie is an urban area I used a search Radius of 1
mile. I opened the
centroid.dbf and saved it as near.csv using excel. I was then able to add this layer to
QGIS using the Add Delimited Text Layer making sure to choose no geometry
before selecting ok. I then
needed to join my near.csv to my Study Area. This was completed using the add
vector join and STFID as the join and target field and made sure to cache join
layer in virtual memory. I
could then use the attribute table to select features using an
expression. Any near
distance equal to -1 was a food desert and by inverting the selection the
remaining items were food oases. I
also need to create some basic statistics using the Food Desert Layer and the
POP2000 as the target field. Once
all my data was organized and symbolized I chose the Basic Style in Mapbox,,
imported my tilesets, created my layers and applied symbolization and other
changes to the basic style. I
recorded the color scheme I wanted to use via ColorBrewer in both HEX and RGB
with an application of 5 classes for later use. Within Mapbox I grouped my food desert
layer and then duplicated until I had 5 copies. I set the parameters for my 5 class
Jenks natural breaks classification (determined in ArcMap) and used the RGB
information from ColorBrewer to delineate the food desert breakdown. I made sure to record my map position
(center lat/long and zoom level) so that I could later use this information in
my leaflet code, both via notepad text and then saved as html utf-8 encoding
when complete. Once all
Mapbox edits were complete I published my mapbox and then obtained the leaflet
URL to add to my leaflet code. I
edited all required information in my leaflet code: title, set view coordinates
and zoom level, map size, geocoder coordinates representing the center of the
city and of course pointing all file paths to the correct location. (Since I work from my C drive they
were locally pathed to the leaflet folder within the Analyze2 folder inside my
Data folder. I also confirmed
that my leaflet_rh.html link worked properly and that my information displayed
correctly. I feel that my
food desert results are fairly accurate as I used the most up to date data
available. Also as local
resident, and city worker privy to development, I was able to determine that
the pattern fit rather well. The
food deserts may have a slight overestimation as we have had recent growth and
several new grocery stores built but all in all I’d give it at a minimum a 90%
rating. This is also due to
Google Earth not having all of the new grocery store locations updated.
2. Describe the data being
represented. Are there trends you can identify? If you used your local town to map,
did you expect to see food deserts occurring where they are? Does the data surprise you?
I used my local
town, the City of Port Saint Lucie. The data illustrated in my leafelet
map shows the location of existing grocery stores (obtained from google earth)
and the location of Food Deserts overlaid on a base map to visualize where they
are located within the city. Because
I live in a well-developed area in the heart of the city with access to many
things, including grocery stores I did not expect to find the number of food
deserts illustrated on the map. Once
I studied the locations it made perfect sense as over the years the city has
annexed more and more properties and outside contractors had begun
development. When the
economy began to fail several years ago, development stopped, including
commercial entities such as grocery stores which were planned in proximity to
many of the developments. I’m happy to say that trend is beginning to
turn so I expect that the food deserts will lesson over the next few
years.
