#ctnamecleaner, ctpopulator, ctcorrelator

---

ctnamecleaner

An R package that finds that takes a list of Connecticut hamlets or neighborhoods and adds a column with the matching official town names.

• Matches hamlets or neighborhoods to equivalent town names
• Adds additional column with population
• Optionally exports a new CSV file
• Lists names that could not be matched to towns
• Based on an ever-growing list of town names that TrendCT.org comes across.

####What function ctnamecleaner() does Let's assume you have a dataframe in R called towncoffeeshops that looks like

Town	Coffeeshops
Andover	2
Centerbrook	5
Yalesville	1

Run this in R

ctnamecleaner(Town, towncoffeeshops, filename="towncoffeecleaned", case="Upper")

You'll get a new file called towncoffeecleaned.csv that looks like

Town	Coffeeshops	real.town.name
Andover	2	ANDOVER
Centerbrook	5	ESSEX
Yalesville	1	WALLINGFORD

Alternatively

ctnamecleaner(Town, towncoffeeshops)

The command above will create a dataframe without exporting.

####Usage

ctnamecleaner(name, data, filename="nope", case="Title")

####Arguments

• name - Column with town names
• data - Name of data frame.
• filename Name of CSV to save. If skipped, CSV will not export.
• case Output of town string. Options are Upper, Lower, and Title

---

ctpopulator

An R package that appends the most-recent population of Connecticut towns to a dataframe for efficient per-capita calculations.

• Matches 2013 town population to equivalent town names
• Adds additional column with population
• Optionally exports a new CSV file
• Works best if the dataframe has originally been checked with ctnamecleaner()

####What function ctpopulator() does Let's assume you've collapsed duplicate town names column real.town.name in the CTNAMECLEANED dataframe above and summed up or averaged the figures you were working with.

Run this in R

ctpopulator(real.town.name, CTNAMECLEANED, filename="towncoffeepop")

You'll get a new file called towncoffeepop.csv that looks like the table below. Note: if you exclude the CSV filename parameter only the dataframe will be exported and can be assigned to an object.

Town	Coffeeshops	real.town.name	pop2013
Andover	2	ANDOVER	3095
Centerbrook	5	ESSEX	6668
Yalesville	1	WALLINGFORD	45112

####Usage

ctnamecleaner(name, data, filename="nope")

####Arguments

• name - Column with town names
• data - Name of data frame.
• filename Name of CSV to save. If skipped, CSV will not export.
• case Output of town string. Options are Upper, Lower, and Title

---

ctcorrelator

An R package that takes a town dataframe and checks for correlations between the original data set and 500 different variables including demographics, median income, education attainment, and poverty from an ever-growing list. Why? Correlation does not mean causation. But having a quickly generated list could help point a researcher of journalist into unforseen directions with respect to the original data.

• Appends the original dataframe with ~500 other variables from ~20 spreadsheets
• Determines the Pearson product-moment correlation coefficient for each variable as compared to the original data
• Notifies the users which variables had the strongest correlations by exporting tables and charts to browse
• Exports a summary shreadsheet _ Exports a dataframe of the variables with strong correlations for analysis later _ Exports small-multiple scatterplots based on user-set coefficient threshold
• Works best if the dataframe has originally been checked with ctnamecleaner()

####What function ctcorrelator() does Let's assume you've collapsed duplicate town names column real.town.name in the CTNAMECLEANED dataframe above and summed up or averaged the figures you were working with.

This is a dataframe called ctcoffeeshops.

Town	Coffeeshops
Andover	2
Essex	5
Wallingford	1

Run this in R

ctcorrelator(ctcoffeeshops, p=.9)

You'll get a new file called array_summary.csv that looks similar to this:

row	correlation	n()
1	moderate.negative.correlation	7
2	moderate.positive.correlation	70
3	no.correlation	12
4	strong.negative.correlation	3
5	strong.positive.correlation	103
6	very.strong.positive.correlation	8
7	weak.negative.correlation	6
8	weak.positive.correlation	49

You'll get a new file called strong.very.strong.csv that looks similar to this:

row	column.abbrev	corre	correlation	raw	column.name
1	below.poverty	0.947982822	very.strong.positive.correlation	0.947982822	Below poverty
2	g11	0.934302408	very.strong.positive.correlation	0.934302408	Educational Attainment for the Population 25 Years and Over, 11th grade (City)
3	female.householder.male.partner	0.931860863	very.strong.positive.correlation	0.931860863	Unmarried-partner Households by Sex of Partner, Female householder and male partner (City)

And then you'll also get a new file called plot.png that looks similar to

####Usage

ctcorrelator(dat_data, p=.9)

####Arguments

• dat_data - dataframe with two columns: town with CT town names and one other one with raw numbers representing whatever you need.
• p - The minimum threshold for the correlation coefficient that will be rendered in the charts.

---

####What you'll need to start

• R
• RStudio (not really, but sure)

####What to run within R or RStudio Assuming user is starting from scratch

install.packages("devtools")
library(devtools)

install_github("trendct/ctnamecleaner")
library(ctnamecleaner)

Future versions

Will account for zip codes and census tracts or possibly blocks in Connecticut.

Version

0.3.1

MIT