Coding in R for Policy Analytics



This lab offers practice with logical statements used to create groups from your data.

I have provided you with a LAB-02 RMD template:



Getting Started

Functions Used in this Lab

You will use the following functions for this lab:

logical operators  ( ==, >, <, ! )
sum()        # count TRUEs in logical vectors
mean()       # proportion of TRUEs in logical vectors
compound statements with & (AND) and | (OR) operators



Data

This lab uses city tax parcel data, and open dataset released by the municipal government of Syracuse, NY.

  • Each house, business, park or empty lot in the city sits on a parcel of land, and the tax rate for that parcel is determined by the city.
  • All of the characteristics of each parcel is stored in a central database, along with information about the land use classifications, owners, and whether the property is up-to-date on taxes and water bills.
  • Each row represents one tax parcel, and all tax parcels in the city are recorded in the dataset.

We will use the downtown area for this lab:



Load the Downtown Syracuse Parcel Map

You will need the following packages for this lab:

install.packages( "geojsonio" )
install.packages( "sp" )
install.packages( "rgdal" )

NOTE: do not include include install package commands in your RMD chunks. Trying to install packages while knitting can cause errors.

library( geojsonio )
library( sp )
library( rgdal )

Load the map file:

URL <- "https://raw.githubusercontent.com/DS4PS/Data-Science-Class/master/DATA/downtown-syr.geojson"
downtown <- geojson_read( URL, what="sp" )
plot( downtown,  border="gray50", col="gray80" )



Data Dictionary

Note, the downtown file contains a subset of the fields.

Table continues below
tax_id stnum stname zip
31150010100000100080010000 484-98 SALINA ST S & ONONDAGA ST 13078
31150010300000150050000000 311-17 WATER ST E TO ERIE BLVD E 13202
31150010300000160010000000 257 WATER ST E & MONTGOMERY S 13202
31150010300000160020000000 251 WATER ST E TO ERIE BLVD E 14534
31150010300000160030000000 243-49 WATER ST E TO ERIE BLVD E 13066
31150010300000160040000000 239 WATER ST E TO ERIE BLVD E 13088
31150010300000160050000000 235 WATER ST E TO ERIE BLVD E 13202
31150010300000160060000000 207-33 WATER ST E TO ERIE BLVD E 13202
31150010300000160070000000 203 WATER ST E & WARREN ST S 13202
31150010400000170010000000 135 WATER ST E & WARREN ST S 13088
Table continues below
frontfeet depth sqft acres nhood ward censustract
143.7 273.2 57935 1.33 Downtown 09 32
79.09 78.71 6209 0.1425 Downtown 15 32
41 60 2475 0.05681 Downtown 09 32
40.89 60 2467 0.05662 Downtown 09 32
41.63 60 2499 0.05738 Downtown 09 32
20.81 60 1244 0.02856 Downtown 09 32
20.81 60 1248 0.02865 Downtown 09 32
136.9 60 8169 0.1875 Downtown 09 32
28.26 59 1694 0.03889 Downtown 09 32
57.3 59.44 3390 0.07783 Downtown 09 32
Table continues below
landuse yearbuilt units vacantbuild watersvc assessedland
Parking 0 0 NA NA 650000
Commercial 1950 0 N A 61600
Parks 0 0 N NA 24600
Commercial 1900 0 N A 24000
Commercial 1950 0 N A 24600
Commercial 1950 0 N A 24000
Commercial 1980 0 N A 24000
Commercial 1980 0 N A 104400
Commercial 1940 0 N A 33000
Commercial 1887 0 N A 100900
Table continues below
assessedval citytaxable starc taxtrust amtdelinqt taxyrsdelinqt
650000 650000 650000 NA 0 0
112000 0 0 NA 0 0
24600 0 0 NA 0 0
211000 211000 211000 NA 0 0
384000 384000 384000 NA 0 0
160000 160000 160000 NA 0 0
166000 166000 166000 NA 0 0
225000 225000 225000 NA 0 0
396000 205000 205000 NA 0 0
3e+05 3e+05 3e+05 NA 0 0
overduewater totint
0 0
0 0
0 0
0 0
0 0
45.64 0
2.13 0
2.41 0
323.9 0
0 0





Lab Instructions

Answer the following questions using the Syracuse downtown parcels dataset and the functions listed.

Your solution should include a written response to the question, as well as the code used to generate the result.

The images are included so that you can check whether your logical statements are correct.


Questions

1. Where are the majority of vacant lots located in the downtown? Produce a map of your results that matches the one shown here.

landuse variable


2. How many parking lots are in downtown Syracuse? Map your results.

sum() function and landuse variable



3. Where is new construction located in the city?

Calculate the number of buildings constructed since 1980.

yearbuilt variable

What proportion of commercial properties are built since 1980?

count commercial properties, and buildings constructed since 1980: landuse and yearbuilt variables


4. How many parcels have assessed values over $10 million? Map your results.

use the sum() function with assessedval variable;


5. How many properties are delinquent on taxes? Map your results.

sum() with the amtdelinqt;


6. What proportion of parcels have delinquent tax payments owed?

mean() with the amtdelinqt variable


7. What proportion of commercial properties are delinquent on taxes? What proportion of delinquent tax bills are owed by commercial parcels?

mean() with amtdelinqt and landuse variables define your groups with compound statements first then calculate proportions


8. How many of each land use type are delinquent on taxes?

table() with amtdelinqt and landuse variables define delinquency first, then create a table with landuse and delinquency



Challenge Questions

Map all of the land valued at over $1 million per acre.

Map all of the land valued at less than $500,000 per acre.

What is the total value of all of the commercial parcels in the city?

What is the total value of all of the non-commercial parcels in the city?


Submission Instructions

When you have completed your assignment, knit your RMD file to generate your rendered HTML file.

Login to Canvas at http://canvas.asu.edu and navigate to the assignments tab in the course repository. Upload your HTML and RMD files to the appropriate lab submission link.

Platforms like BlackBoard and Canvas sometimes disallow you from submitting HTML files when there is embedded computer code. If this happens create a zipped folder with both the RMD and HTML files.

Remember to:

  • name your files according to the convention: Lab-##-LastName.Rmd
  • show your solution, include your code.
  • do not print excessive output (like a full data set).
  • follow appropriate style guidelines (spaces between arguments, etc.).

See Google’s R Style Guide for examples.


Markdown Trouble?

If you are having problems with your RMD file, visit the RMD File Styles and Knitting Tips manual.

Notes on Knitting

Note that when you knit a file, it starts from a blank slate. You might have packages loaded or datasets active on your local machine, so you can run code chunks fine. But when you knit you might get errors that functions cannot be located or datasets don’t exist. Be sure that you have included chunks to load these in your RMD file.

Your RMD file will not knit if you have errors in your code. If you get stuck on a question, just add eval=F to the code chunk and it will be ignored when you knit your file. That way I can give you credit for attempting the question and provide guidance on fixing the problem.



Helpful Hints

Some things to keep in mind as you get started:

Steps in Identifying Groups

For this lab you will construct a group by translating some management question from plain English into a logical statement, then show a map of the newly constructed group by adapting the code provided:

How many parcels are larger than one acre?

these <- downtown$acres > 1   # the group I am defining
sum( these )                  # count of parcels in our defined group
## [1] 59
mean( these )                 # proportion of total parcels in our defined group
## [1] 0.151671

To show the location of this group on a map:

these <- downtown$acres > 1                               # define your group
group.colors <- ifelse( these, "firebrick", "gray80" )    # don't change this
plot( downtown,  border="gray70", col=group.colors )      # don't change this

What proportion of downtown parcels are occupied by commercial buildings?

these <- downtown$landuse == "Commercial"
mean( these )  # 54%
## [1] 0.5372751

Plot commercial parcels on a map:

these <- downtown$landuse == "Commercial"
group.colors <- ifelse( these, "firebrick", "gray90" )     # don't change this
plot( downtown,  border="gray70", col=group.colors )       # don't change this



Identifying Unique Values

We will often define a group using criteria relevant to the management question. For example, we might want to look at all single family homes that are valued over $250,000 in a city. Or perhaps we want to identify all commercial properties that are delinquent on taxes. These are examples of compound logical statements that combine information from two or more variables (land use and value, or land use and tax status).

# single family homes worth more than $250,000
downtown$landuse == "Single Family"   &    downtown$assessedval > 250000

We can find the current category levels defined within factors or character vectors using a table() or unique() function.

table( downtown$landuse )
## 
##          Apartment         Commercial Community Services         Industrial 
##                  6                209                 17                  4 
##            Parking              Parks         Recreation          Religious 
##                 78                  8                  5                  6 
##            Schools      Single Family          Utilities        Vacant Land 
##                  4                  1                  6                 45
unique( downtown$landuse )
##  [1] "Parking"            "Commercial"         "Parks"             
##  [4] "Community Services" "Vacant Land"        "Utilities"         
##  [7] "Apartment"          "Recreation"         "Schools"           
## [10] "Religious"          "Industrial"         "Single Family"

Note that the spelling has to be precise for the statement to work correctly.

sum( downtown$landuse == "Apartment" )
## [1] 6
sum( downtown$landuse == "apartment" )
## [1] 0



Missing Values

When there are missing values (NAs) in a vector you need to tell mathematical functions like sum() and mean() to ignore them, otherwise they will return NA.

x <- c( 1, 2, NA, 4 )
sum( x )
## [1] NA
sum( x, na.rm=T )
## [1] 7

Any ideas why R would default to NA for mathematical functions?

Consider an example with a free museum that brings in a special exhibit. They don’t charge when you enter, but if you visit the special exhibit you have the option to make a donation to support the event. The computer logs all visitors, but it only enters values for those that visited the special exhibit.

x <- c( NA, NA, 20, NA, 0, 15, NA, 45, NA, NA, NA, NA, 10, 0, NA, NA )
# how many people visited the museum today?
length( x )
## [1] 16
# how much was donated to the special exhibit? 
sum( x, na.rm=T )
## [1] 90
# what was the average donation made to the special exhibit? 
mean( x, na.rm=T )
## [1] 15
# if we know that 200 people will visit the meseum this weekend,
# how much do we expect the special exhibit to raise? 
#
# people visiting x average donation
200 * mean( x, na.rm=T )
## [1] 3000

Is this a correct estimate?

What average does this represent (who is in the denominator)?

x <- c( NA, NA, 20, NA, 0, 15, NA, 45, NA, NA, NA, NA, 10, 0, NA, NA )
# what was the average donation made to the special exhibit? 
mean( x, na.rm=T )
## [1] 15

When you add na.rm=T to the function it will drop all cases with missing data.

Let’s look at the data again and think about the counts of people more precisely:

# number of people that visit the museum
length( x )
## [1] 16
# number of people that attended the exhibit
sum( x >= 0, na.rm=T )
## [1] 6
# proportion of museum visitors that attended the exhibit
sum( x >= 0, na.rm=T ) / length( x )
## [1] 0.375
# average donation for ALL museum visitors
sum( x, na.rm=T ) / length( x )
## [1] 5.625
# average donation for ONLY THOSE that went to the exhibit
sum( x, na.rm=T ) / sum( x >= 0, na.rm=T )
## [1] 15

Our initial estimate was incorrect. We multiplied all museum visitors by the average donation of those that attended the exhibit, a subset of all visitors. We will greatly over-estimate the expected event revenue as a result.

# if we know that 200 people will visit the meseum this weekend,
# how much do we expect the special exhibit to raise? 
#
# people visiting x average donation per person
200 * mean( x, na.rm=T )
## [1] 3000

To get an accurate estimate of the special event revenues we need to consider how many museum visitors will attend the exhibit:

# accurate estimate:
# people x those at exhibit x ave donation
200 * ( sum( x >= 0, na.rm=T ) / length( x ) ) * mean( x, na.rm=T )
## [1] 1125
# alternatively: 
# number of visitors x ave donation per museum visitor
200 * ( sum( x, na.rm=T ) / length(x) )
## [1] 1125

Missing values serve a very important function in statistics, because they force the analyst to decide whether we can ignore them completely, or whether the need to be interpretted if we want the model to be accurate. So when you get this:

mean( x )
## [1] NA

R is telling you to take a step back to determine how we should treat these values. We can interpret missing values as zero when we tabulate the total of donations:

sum( x, na.rm=T )
## [1] 90

But it is problematic when we are looking at proportions or averages in this context.

Missing values means something different in every context, so pay attention to them when they exist. I guarantee that you will make an error in an important calculation at least once in your career because you ignore missing values.



Compound Logical Statements

We often want to create a new group by combining existing groups:

public.goods  <-  downtown$landuse == "Parks" | 
                  downtown$landuse == "Parking" | 
                  downtown$landuse == "Vacant Land"
mean( public.goods ) # proportion of downtown
## [1] 0.3367609

The is a compound logical statement because there is more than one criteria used to define the group. Compound statements require the use of the OR operator | or the AND & operator. They can also combine criteria from multiple variables in the model.

For example, a brown-eyed girl would be written:

brown.eyed.girl <- dat$gender == "female" & dat$eye.color == "brown"

Logical statements are one of the most useful tools in data, but they are a little tricky at first. Consider the difference between the criteria FEMALE & BROWN EYES vs FEMALE | BROWN EYES.

In a world where eyes is either brown or green and gender is either male or female, these two criteria would return the following:

FEMALE & BROWN EYES

  • female + brown eyes
  • intersection of all females and all people with brown eyes

FEMALE | BROWN EYES

  • female + brown eyes
  • female + green eyes
  • male + brown eyes
  • all females in the dataset and all people with brown eyes

Note that in plain English we say “females AND people with brown eyes” when we are combining groups, using the OR operator to create a union of groups.

A good way to keep the logic straight is to add the term “ONLY” to the front of each plain English group descriptor. “ONLY females in the dataset AND ONLY people with brown eyes”.

Also pay attention to the order of operations when applying the NOT operator ! to logical statements. The operator flips each TRUE to a FALSE in a logical vector, and vice-versa each FALSE to TRUE. It creates a compliment set, which is anything not in the current group.

Consider these two cases:

! ( FEMALE & BROWN EYES )

  • female + green eyes
  • male + brown eyes
  • male + green eyes
  • the values of brown-eyed girls become FALSE, and everything else is TRUE or part of the new group

! FEMALE & BROWN EYES

  • male + brown eyes
  • first identify NOT females (all males in the data), then intersect that group with brown eyes

Note that the NOT operator is not equivalent to what some people might consider an OPPOSITE operation. The compliment group of brown-eyed girls is not green-eyed boys! That would be written:

! FEMALE & ! BROWN EYES

If you pay attention to the order of operations, just like mathematical operators, it will help you avoid making errors.



How to Submit

Use the following instructions to submit your assignment, which may vary depending on your course’s platform.


Knitting to HTML

When you have completed your assignment, click the “Knit” button to render your .RMD file into a .HTML report.


Special Instructions

Perform the following depending on your course’s platform:

  • Canvas: Upload both your .RMD and .HTML files to the appropriate link
  • Blackboard or iCollege: Compress your .RMD and .HTML files in a .ZIP file and upload to the appropriate link

.HTML files are preferred but not allowed by all platforms.


Before You Submit

Remember to ensure the following before submitting your assignment.

  1. Name your files using this format: Lab-##-LastName.rmd and Lab-##-LastName.html
  2. Show both the solution for your code and write out your answers in the body text
  3. Do not show excessive output; truncate your output, e.g. with function head()
  4. Follow appropriate styling conventions, e.g. spaces after commas, etc.
  5. Above all, ensure that your conventions are consistent

See Google’s R Style Guide for examples of common conventions.



Common Knitting Issues

.RMD files are knit into .HTML and other formats procedural, or line-by-line.

  • An error in code when knitting will halt the process; error messages will tell you the specific line with the error
  • Certain functions like install.packages() or setwd() are bound to cause errors in knitting
  • Altering a dataset or variable in one chunk will affect their use in all later chunks
  • If an object is “not found”, make sure it was created or loaded with library() in a previous chunk

If All Else Fails: If you cannot determine and fix the errors in a code chunk that’s preventing you from knitting your document, add eval = FALSE inside the brackets of {r} at the beginning of a chunk to ensure that R does not attempt to evaluate it, that is: {r eval = FALSE}. This will prevent an erroneous chunk of code from halting the knitting process.