pls.Rmd

# Public Libraries Survey (PLS) {-}

[![License: GPL v3](https://img.shields.io/badge/License-GPLv3-blue.svg)](https://www.gnu.org/licenses/gpl-3.0) <a href="https://github.com/asdfree/pls/actions"><img src="https://github.com/asdfree/pls/actions/workflows/r.yml/badge.svg" alt="Github Actions Badge"></a> 

A comprehensive compilation of administrative information on all public libraries in the United States.

* Two tables, with one record per library system and one record per library building or bookmobile.

* Released annually since 1992.

* Conducted by the [Institute of Museum and Library Services (IMLS)](https://www.imls.gov/), collected by the [Census Bureau](http://www.census.gov/).

---

## Recommended Reading {-}

Two Methodology Documents:

> [Data File Documentation and User's Guide](https://www.imls.gov/sites/default/files/2023-06/2021_pls_data_file_documentation.pdf)

> `README FY #### PLS PUD.txt` included in each [zipped file](https://www.imls.gov/sites/default/files/2023-06/pls_fy2021_csv.zip)

<br>

One Haiku:

```{r}
# census, not survey.
# dewey decimal index
# finger to lips shush
```

---

## Download, Import, Preparation {-}

Download and import the most recent administrative entity csv file:
```{r eval = FALSE , results = "hide" }
this_tf <- tempfile()

csv_url <- "https://www.imls.gov/sites/default/files/2023-06/pls_fy2021_csv.zip"

download.file( csv_url , this_tf, mode = 'wb' )

unzipped_files <- unzip( this_tf , exdir = tempdir() )
		
administrative_entity_csv_fn <-
	unzipped_files[ grepl( 'AE(.*)csv$' , basename( unzipped_files ) ) ]

pls_df <- read.csv( administrative_entity_csv_fn )

names( pls_df ) <- tolower( names( pls_df ) )

pls_df[ , 'one' ] <- 1
```

Recode missing values as described in the readme included with each zipped file:
```{r eval = FALSE , results = "hide" }
for( this_col in names( pls_df ) ){

	if( class( pls_df[ , this_col ] ) == 'character' ){
	
		pls_df[ pls_df[ , this_col ] %in% 'M' , this_col ] <- NA
		
	}
	
	if( 
		( class( pls_df[ , this_col ] ) == 'numeric' ) | 
		( this_col %in% c( 'phone' , 'startdat' , 'enddate' ) ) 
	){
	
		pls_df[ pls_df[ , this_col ] %in% c( -1 , -3 , -4 , -9 ) , this_col ] <- NA
		
	}
	
}
```

### Save Locally \ {-}

Save the object at any point:

```{r eval = FALSE , results = "hide" }
# pls_fn <- file.path( path.expand( "~" ) , "PLS" , "this_file.rds" )
# saveRDS( pls_df , file = pls_fn , compress = FALSE )
```

Load the same object:

```{r eval = FALSE , results = "hide" }
# pls_df <- readRDS( pls_fn )
```

### Variable Recoding {-}

Add new columns to the data set:
```{r eval = FALSE , results = "hide" }
pls_df <- 
	transform( 
		pls_df , 
		
		c_relatn = 
			factor( c_relatn , levels = c( "HQ" , "ME" , "NO" ) ,
				c( "HQ-Headquarters of a federation or cooperative" ,
				"ME-Member of a federation or cooperative" ,
				"NO-Not a member of a federation or cooperative" )
			) ,
			
		more_than_one_librarian = as.numeric( libraria > 1 )
				
	)	
```

---

## Analysis Examples with base R \ {-}

### Unweighted Counts {-}

Count the unweighted number of records in the table, overall and by groups:
```{r eval = FALSE , results = "hide" }
nrow( pls_df )

table( pls_df[ , "stabr" ] , useNA = "always" )
```

### Descriptive Statistics {-}

Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
mean( pls_df[ , "popu_lsa" ] , na.rm = TRUE )

tapply(
	pls_df[ , "popu_lsa" ] ,
	pls_df[ , "stabr" ] ,
	mean ,
	na.rm = TRUE 
)
```

Calculate the distribution of a categorical variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
prop.table( table( pls_df[ , "c_relatn" ] ) )

prop.table(
	table( pls_df[ , c( "c_relatn" , "stabr" ) ] ) ,
	margin = 2
)
```

Calculate the sum of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
sum( pls_df[ , "popu_lsa" ] , na.rm = TRUE )

tapply(
	pls_df[ , "popu_lsa" ] ,
	pls_df[ , "stabr" ] ,
	sum ,
	na.rm = TRUE 
)
```

Calculate the median (50th percentile) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
quantile( pls_df[ , "popu_lsa" ] , 0.5 , na.rm = TRUE )

tapply(
	pls_df[ , "popu_lsa" ] ,
	pls_df[ , "stabr" ] ,
	quantile ,
	0.5 ,
	na.rm = TRUE 
)
```

### Subsetting {-}

Limit your `data.frame` to more than one million annual visits:
```{r eval = FALSE , results = "hide" }
sub_pls_df <- subset( pls_df , visits > 1000000 )
```
Calculate the mean (average) of this subset:
```{r eval = FALSE , results = "hide" }
mean( sub_pls_df[ , "popu_lsa" ] , na.rm = TRUE )
```

### Measures of Uncertainty {-}

Calculate the variance, overall and by groups:
```{r eval = FALSE , results = "hide" }
var( pls_df[ , "popu_lsa" ] , na.rm = TRUE )

tapply(
	pls_df[ , "popu_lsa" ] ,
	pls_df[ , "stabr" ] ,
	var ,
	na.rm = TRUE 
)
```

### Regression Models and Tests of Association {-}

Perform a t-test:
```{r eval = FALSE , results = "hide" }
t.test( popu_lsa ~ more_than_one_librarian , pls_df )
```

Perform a chi-squared test of association:
```{r eval = FALSE , results = "hide" }
this_table <- table( pls_df[ , c( "more_than_one_librarian" , "c_relatn" ) ] )

chisq.test( this_table )
```

Perform a generalized linear model:
```{r eval = FALSE , results = "hide" }
glm_result <- 
	glm( 
		popu_lsa ~ more_than_one_librarian + c_relatn , 
		data = pls_df
	)

summary( glm_result )
```

---

## Replication Example {-}
This example matches Interlibrary Relationship Frequencies on [PDF page 169 of the User's Guide](https://www.imls.gov/sites/default/files/2023-06/2021_pls_data_file_documentation.pdf#page=169):

```{r eval = FALSE , results = "hide" }
# remove closed and temporarily closed libraries
results <- table( pls_df[ !( pls_df[ , 'statstru' ] %in% c( 3 , 23 ) ) , 'c_relatn' ] )

stopifnot( results[ "HQ-Headquarters of a federation or cooperative" ] == 112 )
stopifnot( results[ "ME-Member of a federation or cooperative" ] == 6859 )
stopifnot( results[ "NO-Not a member of a federation or cooperative" ] == 2236 )
```

---

## Analysis Examples with `dplyr` \ {-}

The R `dplyr` library offers an alternative grammar of data manipulation to base R and SQL syntax. [dplyr](https://github.com/tidyverse/dplyr/) offers many verbs, such as `summarize`, `group_by`, and `mutate`, the convenience of pipe-able functions, and the `tidyverse` style of non-standard evaluation. [This vignette](https://cran.r-project.org/web/packages/dplyr/vignettes/dplyr.html) details the available features. As a starting point for PLS users, this code replicates previously-presented examples:

```{r eval = FALSE , results = "hide" }
library(dplyr)
pls_tbl <- as_tibble( pls_df )
```
Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
pls_tbl %>%
	summarize( mean = mean( popu_lsa , na.rm = TRUE ) )

pls_tbl %>%
	group_by( stabr ) %>%
	summarize( mean = mean( popu_lsa , na.rm = TRUE ) )
```

---

## Analysis Examples with `data.table` \ {-}

The R `data.table` library provides a high-performance version of base R's data.frame with syntax and feature enhancements for ease of use, convenience and programming speed. [data.table](https://r-datatable.com) offers concise syntax: fast to type, fast to read, fast speed, memory efficiency, a careful API lifecycle management, an active community, and a rich set of features. [This vignette](https://cran.r-project.org/web/packages/data.table/vignettes/datatable-intro.html) details the available features. As a starting point for PLS users, this code replicates previously-presented examples:

```{r eval = FALSE , results = 'hide' }
library(data.table)
pls_dt <- data.table( pls_df )
```
Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = 'hide' }
pls_dt[ , mean( popu_lsa , na.rm = TRUE ) ]

pls_dt[ , mean( popu_lsa , na.rm = TRUE ) , by = stabr ]
```

---

## Analysis Examples with `duckdb` \ {-}

The R `duckdb` library provides an embedded analytical data management system with support for the Structured Query Language (SQL). [duckdb](https://duckdb.org) offers a simple, feature-rich, fast, and free SQL OLAP management system. [This vignette](https://duckdb.org/docs/api/r) details the available features. As a starting point for PLS users, this code replicates previously-presented examples:

```{r eval = FALSE , results = 'hide' }
library(duckdb)
con <- dbConnect( duckdb::duckdb() , dbdir = 'my-db.duckdb' )
dbWriteTable( con , 'pls' , pls_df )
```
Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = 'hide' }
dbGetQuery( con , 'SELECT AVG( popu_lsa ) FROM pls' )

dbGetQuery(
	con ,
	'SELECT
		stabr ,
		AVG( popu_lsa )
	FROM
		pls
	GROUP BY
		stabr'
)
```