Project 2: Manipulating and Summarizing Attribute Data
By Luminita Slevoaca
In this project we analyze the damages created along the path of the tornadoes that struck Oklahoma on May 3rd, 1999. “A total of 74 tornadoes touched down across Oklahoma and Kansas in less than 21 hours. At one point, there were as many as four tornadoes reported on the ground at the same time. The strongest tornado, rated a maximum F-5 on the Fujita Tornado Scale, tracked for nearly an hour and a half along a 38-mile path from Chickasha through south Oklahoma City and the suburbs of Bridge Creek, Newcastle, Moore, Midwest City and Del City. As the skies cleared, the states counted 46 dead and 800 injured, more than 8000 homes damaged or destroyed, and total property damage of nearly $1.5 billion”. (NOAA National Severe Storm Laboratory, 2009)
Initial data
For our analysis we have been provided with two shapefiles, “Counties” and “TornadoPaths”, and three attribute tables, “county_damage”, “county_statistics” and “tornado_attributes”. The shapefiles were provided in geographic coordinates and a projection file was not included in the download. A projection is necessary for the data sets to be aligned properly and if performing calculations of distance or area, for example, to yield accurate results. The first step was to open the shapefiles in ArcCatalog and to define a projection using the “Define”Projection” tool from ArcToolbox. The “Define”Projection” tool changed the metadata of our shapefiles from “GCS_Assumed_Geographic _1” to “GCS_North_American_1983”. Defining a coordinate system for our shapefiles allows the software to project the data on the fly. (Zeiders, personal communication, 2010) The next step was to project our data, since it was expressed in unprojected geographic coordinates. To perform the projection we used the “Project” tool from ArcToolbox and selected Oklahoma North State Plane Coordinates -Zone 3501 with map units in feet.
Attribute Tables
In this project we were introduced to attribute tables. Attribute tables contain information about associated geographic features. There are two operations that can link the information from the attribute table to the geographic features in the shapefiles: join and relate. Join is used for one-to-one and many-to-one relationships, meaning that one or may features in the first table is associated to one feature from the second table. Relate is used for one-to-many or many-to-many relationships. (King, 2010)
Besides performing joins on the attribute tables, we learned in this lesson to add new fields to the tables and to use the “Field Calculator” to populate them. We used the “Field Calculator” to concatenate character strings and also to perform various numerical calculations.
Analysis
As mentioned before, we were provided with attribute tables that contain statistical information about the housing units that were damaged or destroyed by county during the May 3, 1999 tornado. To create a thematic map of the housing units that were damaged, we first joined the “county_damage” table to a shapefile that represents Oklahoma’s counties. In order add the “county_statistics” table to the same shapefile, we needed a common field to base the join operation on. To realize that, we created a new field “St_CNTY_FP”, that we populated using the “Field Calculator” to concatenate the “STATE_FIPS” and the “CNTY_FIPS” strings. Then, we performed “Select by Attribute” to select only the counties that were affected by the tornadoes. It resulted that 17 counties were affected and we exported our selection into a new shapefile. We added new fields in the attribute table associated damaged counties shapefile: “Area_sqmi”, “hsng_units”, “tot_damage”, “tot_destr”,”tot_dollar” and “hsng_dens”. We populated these fields using “Calculate Geometry” for “Area_sqmi” and the “Field Calculator” for the rest of the fields.
To create a cloropleth map of the damaged properties, I used a Jenks Natural Breaks classification with 5 classes. The “F Scale” layer present in the legend represents the tornado paths and is color-coded by intensity according to Fujita Tornado Scale, where 1 is the minimum intensity and 5 is the maximum.
By Luminita Slevoaca
In this project we analyze the damages created along the path of the tornadoes that struck Oklahoma on May 3rd, 1999. “A total of 74 tornadoes touched down across Oklahoma and Kansas in less than 21 hours. At one point, there were as many as four tornadoes reported on the ground at the same time. The strongest tornado, rated a maximum F-5 on the Fujita Tornado Scale, tracked for nearly an hour and a half along a 38-mile path from Chickasha through south Oklahoma City and the suburbs of Bridge Creek, Newcastle, Moore, Midwest City and Del City. As the skies cleared, the states counted 46 dead and 800 injured, more than 8000 homes damaged or destroyed, and total property damage of nearly $1.5 billion”. (NOAA National Severe Storm Laboratory, 2009)
Initial data
For our analysis we have been provided with two shapefiles, “Counties” and “TornadoPaths”, and three attribute tables, “county_damage”, “county_statistics” and “tornado_attributes”. The shapefiles were provided in geographic coordinates and a projection file was not included in the download. A projection is necessary for the data sets to be aligned properly and if performing calculations of distance or area, for example, to yield accurate results. The first step was to open the shapefiles in ArcCatalog and to define a projection using the “Define”Projection” tool from ArcToolbox. The “Define”Projection” tool changed the metadata of our shapefiles from “GCS_Assumed_Geographic _1” to “GCS_North_American_1983”. Defining a coordinate system for our shapefiles allows the software to project the data on the fly. (Zeiders, personal communication, 2010) The next step was to project our data, since it was expressed in unprojected geographic coordinates. To perform the projection we used the “Project” tool from ArcToolbox and selected Oklahoma North State Plane Coordinates -Zone 3501 with map units in feet.
Attribute Tables
In this project we were introduced to attribute tables. Attribute tables contain information about associated geographic features. There are two operations that can link the information from the attribute table to the geographic features in the shapefiles: join and relate. Join is used for one-to-one and many-to-one relationships, meaning that one or may features in the first table is associated to one feature from the second table. Relate is used for one-to-many or many-to-many relationships. (King, 2010)
Besides performing joins on the attribute tables, we learned in this lesson to add new fields to the tables and to use the “Field Calculator” to populate them. We used the “Field Calculator” to concatenate character strings and also to perform various numerical calculations.
Analysis
As mentioned before, we were provided with attribute tables that contain statistical information about the housing units that were damaged or destroyed by county during the May 3, 1999 tornado. To create a thematic map of the housing units that were damaged, we first joined the “county_damage” table to a shapefile that represents Oklahoma’s counties. In order add the “county_statistics” table to the same shapefile, we needed a common field to base the join operation on. To realize that, we created a new field “St_CNTY_FP”, that we populated using the “Field Calculator” to concatenate the “STATE_FIPS” and the “CNTY_FIPS” strings. Then, we performed “Select by Attribute” to select only the counties that were affected by the tornadoes. It resulted that 17 counties were affected and we exported our selection into a new shapefile. We added new fields in the attribute table associated damaged counties shapefile: “Area_sqmi”, “hsng_units”, “tot_damage”, “tot_destr”,”tot_dollar” and “hsng_dens”. We populated these fields using “Calculate Geometry” for “Area_sqmi” and the “Field Calculator” for the rest of the fields.
To create a cloropleth map of the damaged properties, I used a Jenks Natural Breaks classification with 5 classes. The “F Scale” layer present in the legend represents the tornado paths and is color-coded by intensity according to Fujita Tornado Scale, where 1 is the minimum intensity and 5 is the maximum.
Click on the image to enlarge.
To represent the destroyed properties, I used again a Jenks Natural Breaks classification with 5 classes to create a thematic display. The same classification was used to represent below the total cost in dollar for housing units destroyed and the housing density.
To represent the destroyed properties, I used again a Jenks Natural Breaks classification with 5 classes to create a thematic display. The same classification was used to represent below the total cost in dollar for housing units destroyed and the housing density.
Click on the image to enlarge.
From the above maps we noticed that Oklahoma County, followed by Cleveland County, that were situated of the path of a tornado of maximum intensity F5, were the ones that were the most affected.
From the above maps we noticed that Oklahoma County, followed by Cleveland County, that were situated of the path of a tornado of maximum intensity F5, were the ones that were the most affected.
Click on the image to enlarge.
Click on the image to enlarge.
In the “Try this!” section we have been provided with shapefile that represents the census tracts and “demography” attribute table. Based on this data, I created a housing density map by census tract.
In the “Try this!” section we have been provided with shapefile that represents the census tracts and “demography” attribute table. Based on this data, I created a housing density map by census tract.
Click on the image to enlarge.
To create the above map, I used this time a 5 quantile classification scheme, which contains an equal number of features in each class. I chose the quantile classification for this map because it creates a better spatial distribution pattern. Comparing the housing density map per county (Figure 4) with the housing density map by census tract (Figure 5), we notice that the latest provides added detail regarding the spatial distribution relative to the tornado path. Also, looking at the legends of the two maps, we noticed that the housing density per census tract map has larger density numbers since it covers smaller areas. Below is a close-up of the Oklahoma County that has the highest density per census tract along the path of the F5 tornado.
To create the above map, I used this time a 5 quantile classification scheme, which contains an equal number of features in each class. I chose the quantile classification for this map because it creates a better spatial distribution pattern. Comparing the housing density map per county (Figure 4) with the housing density map by census tract (Figure 5), we notice that the latest provides added detail regarding the spatial distribution relative to the tornado path. Also, looking at the legends of the two maps, we noticed that the housing density per census tract map has larger density numbers since it covers smaller areas. Below is a close-up of the Oklahoma County that has the highest density per census tract along the path of the F5 tornado.
Click on the image to enlarge.
In this lesson I learned how to define and project a shapefile, to join tables, to add new fields to the attribute tables and to use the “Field Calculator” to populate then. I also learned how to create aliases for a field, which is a more user-friendly description of the content of the field. Aliases can be created by accessing the “Properties” of the field.
In this lesson I learned how to define and project a shapefile, to join tables, to add new fields to the attribute tables and to use the “Field Calculator” to populate then. I also learned how to create aliases for a field, which is a more user-friendly description of the content of the field. Aliases can be created by accessing the “Properties” of the field.
Click on the image to enlarge.
I also learned how to summarize and find statistics on a column in an attribute table. To access both the “Summarize” and the “Statistics” option, I right click on the column of interest. While the “Statistics” option will give us all the information like count, minimum, maximum, sum etc. with no additional user input, for the “Summarize” option we have to select summary statistics that we want to include in the output table.
Summary
The thematic maps created for this project are useful tool to visualize the damages caused by the tornadoes on May 3rd, 1999 in Oklahoma. Unfortunately, the path of the most intense tornado coincided with the county that had the highest housing density and the damages are proportional.
Source
King, B., & Zeiders, M. (2009). Problem-Solving with GIS, Lesson 2, Part II. The Pennsylvania State University, World Campus. Retrieved January 27, 2010.
NOAA National Severe Storm Laboratory (2009)– May 3, 1999 Oklahoma/ Kansas Tornado Outbreak. . Retrieved January 27, 2010
I also learned how to summarize and find statistics on a column in an attribute table. To access both the “Summarize” and the “Statistics” option, I right click on the column of interest. While the “Statistics” option will give us all the information like count, minimum, maximum, sum etc. with no additional user input, for the “Summarize” option we have to select summary statistics that we want to include in the output table.
Summary
The thematic maps created for this project are useful tool to visualize the damages caused by the tornadoes on May 3rd, 1999 in Oklahoma. Unfortunately, the path of the most intense tornado coincided with the county that had the highest housing density and the damages are proportional.
Source
King, B., & Zeiders, M. (2009). Problem-Solving with GIS, Lesson 2, Part II. The Pennsylvania State University, World Campus. Retrieved January 27, 2010.
NOAA National Severe Storm Laboratory (2009)– May 3, 1999 Oklahoma/ Kansas Tornado Outbreak. . Retrieved January 27, 2010