GIS Assignment 4- 3/3/2014
For this project I am exploring the walkability of Camden, Maine, particularly in regards to important public services. In order to determine how walkable Camden, Maine is, the quality of the data for city streets, hydrography, schools, fire departments, police stations, and libraries was assessed. This project is focused on walking distance, assuming that a quarter mile is reasonable for a person to walk to access services. For this reason, the positional accuracy of the data must be +/- 100 feet. If the roads or public services are less accurate than this, a person could easily wind up walking farther than expected. There are a few required attributes that are necessary for this project. It is important that street names and names of services are correct so that people can follow the correct directions and find the service that they are looking for. Also, speed limits of roads would be important for making decisions about which roads are the most safe to walk on. It is also crucial that the type of road or water source is clear so that individuals know if it is possible to walk over or on them. For example, there are many private roads which individuals might not want to trespass on to access services. It is important that the data is up to date so that locations are correct and recently created or remodeled roads are displayed. Finally, with this walking project is it is particularly important that all of the roads connect so that a person can tell where they will be able to walk and where the roads are intersected by buildings, forests, water, etc.
For this project it quickly becomes evident that the Maine Department of Transportation roads should be used instead of the Census roads. This is true for a variety of different reasons. To start off, here is a zoomed out image of the two road data sets (the blue lines are Maine DOT data and the pink lines are Tiger data):
From this image it clear that the lines are both fairly accurate when compared to the aerial imagery. To illustrate this accuracy more clearly below are two images.
In the first image (above) the TIGER roads (pink) are following the center line of the aerial image road and we can see that the Maine DOT roads (blue) are approximately 21 feet off. This is a distance that is far smaller than the +/- 100 feet accuracy that we are hoping for in this project.
In the second image (above) the Maine DOT roads perfectly follow the centerline of the aerial image roads, and the Tiger roads are approximately 20 feet off, again well within our hopes for accuracy for this project.
The attribute information varies with these two data sets. While the Tiger data offers street names, a very important aspect of walkability, the Maine DOT data has important attributes like the speed limits on roads (important for safety) and the types of roads, which is important because there are many private roads or drives that individuals would not want to trespass on. However, I do think that the Tiger road names would be extremely important to have.
If we were simply basing which data we would use on accuracy or attributes, both of these data sources would suffice for positional accuracy and Tiger would be best for attributes because of the road names. However, when it comes to currency and completeness, the Maine DOT data is far superior. While the TIGER data is from the 2013 Census, the Maine DOT roads were gathered on 2/25/2014, making them more current and more likely to be up to date with all potential construction projects or moves, and therefore better for my project in terms of currency. The completeness of the Maine DOT data is really what makes it stand apart from the Tiger Data. If we look at the zoomed out image of the both the roads again (below) it is clear that there are a number of place where the Tiger data (pink) is unconnected and incomplete
If we zoom in on one of these sections it becomes clear just how large these expanses of missing road are and how much confusion this could cause for individuals attempting to walk to necessary services.
In this one example (above) there are 143 feet of road, and main road I might add, that exists in real life and in the aerial image but is entirely missing from this data set. Because of this massive incompleteness in the Tiger data set, the Maine DOT data set would be best for my project in terms of completeness.
For data on the hydrography of this region I used NHD data and Tiger data.
NHD Data (above)
Tiger Data (above)
The positional accuracy for these two data sets appear similar aside from the river. The image below is from the Tiger data, but it is the same for the NHD data. It shows 108 feet of what is actually land, according to the aerial image, being labeled as water. This is problematic because a person might think that they cannot walk there, when in reality they can.
The image below is also of the Tiger data, but again the shape is the same in the NHD data, and shows 114 feet that is labeled as not having water, but which actually does have water. an unsuspecting walker could easily find themselves in the ocean because of this 100 feet of flawed data.
For positional accuracy of the ocean, both data sets appear similar, occasionally misrepresenting more than 100 feet of space, which is not up to the positional accuracy standards we set for this project. However, it is important to note that while the Tiger data has a scale of 1:24,000, meaning that the data is accurate to +/- 40 feet, the NHD data has a scale of 1:100,000 which means it is only accurate to +/-167 feet. For this reason the Tiger data would be better for positional accuracy. This is a more complicated decision though because if we look at a major river, it becomes clear that if I have to chose one data set for my project based on positional accuracy, perhaps I should use the NHD data.
Above is an image of the NHD data. The shapefile of the river does occasionally misrepresent the river (for example here 59 feet of what is actually land is deemed water).
However, this slightly inaccurate shapefile is better than the line data from Tiger displayed in the image above. Here a span of 168 feet of water is not clearly identified. For someone who is walking, it is extremely important to know the shape and full coverage of this body of water (which NHD offers) versus only having the Tiger line data.
While the Tiger data is from 2013, the NHD data does not have a date attached to it. For this reason the Tiger data would be best in terms of currency for my project. As previously stated, the NHD or the Tiger data would be best for positional accuracy depending on if I am more concerned about the shape of rivers or about the scale and potential number of feet the data might be off. The attribute information on both of these data sets are similar for the needs of my project. Neither of these data sets has attributes that are particularly important for this project. Finally, both of the data sets appear equally complete. There are no major gaps or missing rivers. This means that they would both be adequate for my project in terms of completeness.
4. Optional layers
The data for the services that an individual might want to walk to was very accurate. Below are descriptions of the accuracy of data for schools, the police station, the library, and the firestation.
The positional accuracy of this data is very good. According to the metadata the scale is 1:5,000, meaning that the data is accurate to a little more than +/- 13.33 feet. The points representing the schools (orange triangles) are located basically in the center of the buildings. While this data is very accurate, it has a small chance of being slightly misleading because there is just a single point to represent an entire building with various entrances. With the Community School and Camden Rockport Middle School the points are right at the front entrances, but with the Seton school the point is in the middle of the building, which could cause confusion if a person is trying to find the front entrance. This data set, however, is not complete. I was particularly impressed that it included two alternative education sites. The Community School is an alternative education program for high school students and the Seton School is for younger Special Education students. However, it is entirely missing one elementary school. The school that is missing is a school that used to be public but traded buildings with a Montessori school three years ago, and now houses this private institution. I know this because I grew up in the town next to this one and am very familiar with the services in downtown Camden. This data is from August 1st, 2012. This is an adequate time period because the schools have not moved around since 2012. The attribute accuracy is adequate for my project needs. The attribute table has the correct names and addresses of the schools.
The police station data point (blue triangle) is very accurate. According to the metadata the scale is 1:5,000, meaning that the data is accurate to a little more than +/- 13.33 feet. It is placed where the fire station ends and the police station begins (they are housed in the same building). However, again we have only a point representing an entire building which could lead to confusion about where an individual can actually access an entrance to the building. However, knowing the building and viewing it on the basemap, I think it is small enough that this would not be too big of an accuracy issue. This data set is complete because there is only one police station in Camden and it is at this location. I know because I have been in this area numerous times. This data set was updated February 26th, 2014 which was extremely recently and is certainly adequate. The attributes are not adequate for this data set. The name of the police station is correct, however, the address is not listed which could be extremely problematic if someone is trying to locate this service.
The position of the library data is very accurate. According to the metadata the scale is 1:5,000, meaning that the data is accurate to a little more than +/- 13.33 feet. Furthermore the point (purple triangle) is placed right next to the entrance of the library. Again, we run into the issue of a single point representing an entire building. Furthermore, this library has a second entrance down below (see red arrow in image) which is not represented on this map and is the most commonly used entrance. This is a complete data set because that is the only library in Camden. I know because I am very familiar with the area. This data set was last updated on August 1st, 2012 which is adequate because the library has not undergone any changes in location since then. The attribute accuracy is adequate for this project because the name of the library and the address of the library are included and correct.
The fire station has positional accuracy because, again, the scale is 1:5,000, suggesting that the data is accurate to a little more than +/- 13.33 feet. The point which represents the fire station (red triangle) is located by the entrance to the fire station. There is a slight issue with the fact that it is a single point rather than a building, because there might be slight confusion about the entrance, but aside from that the positional accuracy is good. This is the only fire station in Camden, which I know because I live in the town over, so the data set is complete. The data set was updated on February 26th, 2014 which is extremely recently and absolutely adequate. The attributes are not adequately accurate because although the name of the fire department is provided and correct, the address is not listed.