rcquan

###############################################
##
##    Attempt no 2 at building a shiny web app
##    for AB Testing use - using global.R
##
##    global.R - loading and defining variables for the global environment
##
###############################################

# Pallette used in some charts as a general indicator color for better or worse that the control group

rcquan

#############################
# Ryan Quan
# GitHub Username: rcquan
# Twitter: @ryancquan
# Email: ryan.quan08@gmail.com
#
# The following code loads the NYC Playgrounds Data and
# shapefiles 
#############################

rcquan

library(ggplot2)
library(scales)
library(dplyr)
library(grid)

hoursToMinutes <- function(hours) {
    # converts character HH:MM data to total minutes
    # Args: 
    #   hours - a character vector in "HH:MM" format
    # Returns:

rcquan

#############################
# Ryan Quan
# NLP and classifcation of the 
# Federalist corpus 
#############################
library(tm)
library(glmnet)
library(infotheo)
library(rpart)
library(ggplot2)

rcquan

library(rCharts)
library(zoo)

# Data transformations
names(medMis_plot) <- c("Month", "MedicalMisconductCount")
medMis_plot$Month <- as.yearmon(medMis_plot$Month)
medMis_plot <- transform(medMis_plot, Month = as.character(Month))

# Morris Line Plot
m1 <- mPlot(x = "Month", y = colnames(medMis_plot)[2], data = medMis_plot, type = "Line")

rcquan

# Time-series plot
library(ggplot2)
library(scales)

# By month
m = ggplot(data = medMis_month, aes(x = Time, y = Freq)) + geom_line()
m = m + scale_x_datetime(labels = date_format("%Y"), breaks = date_breaks("2 years"))
m = m + ylab("No. of Physicians Disciplined") + 
        ggtitle("Physicians Disciplined by NYS DOH from 1990-2013")
m

rcquan

# Convert variable from factor into "Date" class
medMis$Effective.Date <- as.character(medMis$Effective.Date)
medMis$Effective.Date <- as.POSIXct(medMis$Effective.Date, format = "%m/%d/%Y")

# Order dataset by Effective.Date from 1990-2014
medMis <- (medMis[order(medMis$Effective.Date),])

# Create time-series data 
medMis_month <- as.data.frame(table(format(medMis$Effective.Date, "%Y-%m-01")), stringsAsFactors = FALSE)
medMis_plot <- as.data.frame(table(format(medMis$Effective.Date, "%Y-%m")), stringsAsFactors = FALSE)

rcquan

# Check if file exists in working directory; if not, downloads
fileName = "medMis.csv"
if (!file.exists(fileName)){
        url <- "https://health.data.ny.gov/api/views/ebmi-8ctw/rows.csv?accessType=DOWNLOAD"
        fileName <- "medMis.csv"
        download.file(url, fileName, method = "curl")
}
medMis <- read.csv(fileName)

rcquan

library(RCurl)
library(XML)

escapeAmpersands <- function(url) {

	doc <- getURL(url)
	## escape ampersands in XML
	validDoc <- gsub("&", "&amp;", doc)
	validDoc
}

rcquan

library(BiSeq)

readBedgraph <- function(inputDirectory, nSamples = 8) {

	## get list of file names
	fileNames <- list.files(path = inputDirectory, pattern = ".bedGraph")
	## parse sample name from file name
    	sampleNames <- sapply(fileNames,function(fileName) {
				gsub("_r123_R1_val_1.fq_bismark_bt2_pe.bedGraph", "", fileName)
				}, USE.NAMES = FALSE)