Blood Pressure Database Example

Tasks covered:

Working with .RData files [save and load]

Creating and working with functions

Working with date data types

Working with date sequences

Plotting multiples data series on a single chart

Annotating a chart axis with dates and and displaying the dates perpendicular to the axis

Project script: BP_database_example.R

This example R script demonstrates how to use an R data frame as a database. It includes code to create an initial blood pressure database. and two functions, one to add new blood pressure entries to the database, and another to chart the data in the database. Code to create two example 52-week blood pressure databases is also included to help demonstrate the use of the two included functions.

This example demonstrates several interesting skills with R. The code shows how to use dates in a data frame, how to show dates along a chart axis, how to use date data to position a chart legend, how to plot multiple lines [data and reference] on the same chart, and the use of color, line type, and line width in a chart.

Create an initial blood pressure database

The first block of code can be used to create an initial entry in your database.

##########################################################################

# use this code to create your own bp database - skip this if you are simply going to run the example

# add new data entries with add_bp_entry( )

# chart the database with view_bp_chart( )

# create the blood pressure database as a data frame [use this code to create a new database]

bp_db <- as.data.frame(cbind(0, 0, 0)) # create the initial database entry

bp_db[,1] <- Sys.Date() # insert the current date into the first column

names(bp_db) <- c('Date','Systolic', 'Diastolic') # assign column titles

save(bp_db, file = 'bp_db.RData') # save the database

# end of the creation your initial database

##########################################################################

The code creates an initial entry in a data frame containing zeros for all of the values. The current date is inserted into the first column. Column titles are assigned to the data frame, and the data frame is saved as an .RData file. This data frame can be used as a blood pressure database with the functions add_bp_entry( ) and view_bp_chart( ).

Example blood pressure databases

The next block of code creates two example blood pressure databases. These are helpful in testing the functionality of the two project functions that will be discussed below.

##########################################################################
# use this code to create the example bp databases
# create the example databases [these are used in the example below]
# add new data entries with add_bp_entry( )
# chart the database with view_bp_chart( )

# create example data
dateSeq <- seq(as.Date("2017-06-03"), by = "week", length.out = 52) # a sequence of 52 weekly dates
sysSeq <- seq(120, 138) # a set of systolic bp values
diaSeq <- seq(68, 85) # a set of distolic bp values

# a function to create our example bp sequences
sampleSeq <- function(s, x) {
r <- c() # an empty list
for (i in 1:x) { # repeat x times
r <-rbind(r, sample(s, 1)) # add a random value from s to list r
}
r # return list r with x elements
}

# create bob's bp database [the dates will be inserted incorrectly]
# generate example bp values
s <- sampleSeq(sysSeq, 52) # a sequence of 52 random systolic bp values
d <- sampleSeq(diaSeq, 52) # a sequence of 52 random diastolic bp values
bob_bp_db <- as.data.frame(cbind(dateSeq,s,d)) # combine the data into a dataframe
# cbind( ) coerces the dates to integer values
bob_bp_db[,1] <- dateSeq # fix the dates in the first column
names(bob_bp_db) <- c('Date','Systolic','Diastolic') # assign column titles
save(bob_bp_db, file = 'bob_bp_db.RData') # save the database

# create carol's bp database [the dates will be inserted incorrectly]
# generate example bp values
s <- sampleSeq(sysSeq, 52) # a sequence of 52 random systolic bp values
d <- sampleSeq(diaSeq, 52) # a sequence of 52 random diastolic bp values
carol_bp_db <- as.data.frame(cbind(dateSeq,s,d)) # combine the data into a dataframe
# cbind( ) coerces the dates to integer values
carol_bp_db[,1] <- dateSeq # fix the dates in the first column
names(carol_bp_db) <- c('Date','Systolic','Diastolic') # assign column titles
save(carol_bp_db, file = 'carol_bp_db.RData') # save the database
# end of the creation the example databases
###########################################################################

The first three lines of the code generate three sequences: a weekly date sequence spanning 52 weeks, a sequence of systolic blood pressure values from 120 to 138, and a sequence of diastolic blood pressure values from 68 to 85.

# create example data
dateSeq <- seq(as.Date("2017-06-03"), by = "week", length.out = 52) # a sequence of 52 weekly dates
sysSeq <- seq(120, 138) # a set of systolic bp values
diaSeq <- seq(68, 85) # a set of distolic bp values

The next section of code defines a function that we will use to generate the series of systolic and diastolic values for each weekly reading in our example databases.

# a function to create our example bp sequences
sampleSeq <- function(s, x) {
r <- c() # an empty list
for (i in 1:x) { # repeat x times
r <-rbind(r, sample(s, 1)) # add a random value from s to list r
}
r # return list r with x elements
}

The function sampleSeq has two input arguments: a sequence of values and the desired number of values in the output list. The first line of code creates an empty list, names r, where the generated values will be stored. Then the function loops for x iterations. In each loop, the input sequence is randomly sampled for and value and that value is inserted in the output list. After the loop is complete, the output list r of x elements is returned.

The next two code blocks create the two example blood pressure databases. Both blocks are identical with the exception that one creates a database called bob_bp_db and the other creates a database called carol_bp_db, therefore only the first block will be described here.

# create bob's bp database [the dates will be inserted incorrectly]
# generate example bp values
s <- sampleSeq(sysSeq, 52) # a sequence of 52 random systolic bp values
d <- sampleSeq(diaSeq, 52) # a sequence of 52 random diastolic bp values
bob_bp_db <- as.data.frame(cbind(dateSeq,s,d)) # combine the data into a dataframe
# cbind( ) coerces the dates to integer values
bob_bp_db[,1] <- dateSeq # fix the dates in the first column
names(bob_bp_db) <- c('Date','Systolic','Diastolic') # assign column titles
save(bob_bp_db, file = 'bob_bp_db.RData') # save the database

The function sampleSeq( ) is used to generate two 52 element sequences, one for the systolic blood pressure values and another for the diastolic values. These two sequences are combined with the weekly dates into the data frame bob_bp_db using cbind( ). Unfortunately, cbind( ) coerces the date values into integer values before as.data.frame( ) creates our data frame, so the date sequence is reassigned to the first column of bob_bp_db. Now, the first column of the data frame is populated correctly with our sequence of weekly dates. The column titles of Date, Systolic, and Diastolic are added to the data frame. finally the data frame is saved as an RData file. This process is repeated for carol_bp_db.

Manage the blood pressure databases [add and view the data]

The next block of code creates the two functions for the blood pressure database. The function add_bp_entry( ) adds a new entry [current date, systolic BP, diastolic BP] to the database and saves it. The function view_bp_chart( ) charts the data in a selected database.

##########################################################################
# the bp database functions
add_bp_entry <- function(DB, systolic, diastolic) { # add an entry to the db
if(DB == 'bob') {
    load('bob_bp_db.RData') # open the database
    new_row <- as.data.frame(cbind(0, systolic, diastolic)) # create a new row
    new_row[,1] <- Sys.Date() # insert the current date in the first column
    names(new_row) <- c('Date','Systolic', 'Diastolic') # add column titles
    result <- rbind(bob_bp_db, new_row) # add the new row to the database
    bob_bp_db <- result
    save(bob_bp_db, file = 'bob_bp_db.RData') # save the database
} else {
    load('carol_bp_db.RData') # open the database
    new_row <- as.data.frame(cbind(0, systolic, diastolic)) # createa a new row
    new_row[,1] <- Sys.Date() # insert the current date in the first column
    names(new_row) <- c('Date','Systolic', 'Diastolic') # add column titles
    result <- rbind(carol_bp_db, new_row) # add the new row to the database
    carol_bp_db <- result
    save(carol_bp_db, file = 'carol_bp_db.RData') # save the database
}
}

view_bp_chart <- function(DB) { # chart the bp data
if(DB == "bob") {
    load('bob_bp_db.RData') # open the database
    startDate <- bob_bp_db[1,1] # get the first date in the database to position the legend correctly
    # plot a line chart of the systolic bp [note that xaxt='n' supresses the x-axis values]
    plot(bob_bp_db$Systolic ~ bob_bp_db$Date, col='red', type = 'l', ylim = c(50,145), xaxt='n',

        main = 'Bob\'s Blood Pressure', xlab = 'Date', ylab = 'BP')
    # add dates as the x-axis values perpendicular [las = 2] to the axis
    axis(1, bob_bp_db$Date, labels = format(bob_bp_db$Date, "%b-%d"), las = 2, cex.axis = .7)
    # add the diastolic bp line
    lines(bob_bp_db$Diastolic ~ bob_bp_db$Date, col = 'green')
} else {
    load('carol_bp_db.RData') # open the database
    startDate <- carol_bp_db[1,1] # get the first date in the database to position the legend correctly
    # plot a line chart of the systolic bp [note that xaxt='n' supresses the x-axis values]
    plot(carol_bp_db$Systolic ~ carol_bp_db$Date, col='red', type = 'l', ylim = c(50,145), xaxt='n',

        main = 'Carol\'s Blood Pressure', xlab = 'Date', ylab = 'BP')
    # add dates as the x-axis values perpendicular [las = 2] to the axis
    axis(1, carol_bp_db$Date, labels = format(carol_bp_db$Date, "%b-%d"), las = 2, cex.axis = .7)
    # add the diastolic bp line
    lines(carol_bp_db$Diastolic ~ carol_bp_db$Date, col = 'green')
}
grid() # show a grid
abline(h = 120, lty = 3, col = 'red') # normal systolic bp line
abline(h = 140, lty = 3, lwd = 2, col = 'red') # borderline systolic bp line
abline(h = 80, lty = 3, col = 'green') # normal diastolic bp line
abline(h = 90, lty = 3, lwd = 2, col = 'green') # borderline diastolic bp line
# add a legend
legend(as.Date(startDate),115, c('Systolic','Diastolic'),lty = c(1,1), col = c('red','green'))
}
# end of the database functions
##########################################################################

The function add_bp_entry( ) will add a daily entry to the selected blood pressure database.

add_bp_entry <- function(DB, systolic, diastolic) { # add an entry to the db
if(DB == 'bob') {
    load('bob_bp_db.RData') # open the database
    new_row <- as.data.frame(cbind(0, systolic, diastolic)) # create a new row
    new_row[,1] <- Sys.Date() # insert the current date in the first column
    names(new_row) <- c('Date','Systolic', 'Diastolic') # add column titles
    result <- rbind(bob_bp_db, new_row) # add the new row to the database
    bob_bp_db <- result
    save(bob_bp_db, file = 'bob_bp_db.RData') # save the database
} else {
    load('carol_bp_db.RData') # open the database
    new_row <- as.data.frame(cbind(0, systolic, diastolic)) # createa a new row
    new_row[,1] <- Sys.Date() # insert the current date in the first column
    names(new_row) <- c('Date','Systolic', 'Diastolic') # add column titles
    result <- rbind(carol_bp_db, new_row) # add the new row to the database
    carol_bp_db <- result
    save(carol_bp_db, file = 'carol_bp_db.RData') # save the database
}
}

This function has three input arguments: the database to add the new entry to, the systolic blood pressure value, and the diastolic blood pressure value. The function is separated into two duplicate sections. The first section adds an entry into bob_bp_db , and the second section adds an entry into carol_bp_db . Since the second section is essentially a mirror of the first section, so this description will only cover the functionality of the first section.

The first code line opens the database file for editing in this function. Once add_bp_entry( ) ends, the blood pressure database is not active. The next line of code, a new data frame is created containing 0, the input systolic value, and the input diastolic value using as.data.frame( ) and cbind( ) . The first value of this data frame is replaced with the current date using Sys.Date( ) and the column titles [Date, Systolic, and Diastolic] are added with the function names( ) . It is necessary to have matching column titles so that new new data frame can be added to the existing database. R will raise an error if the column titles do not match. Finally, the new row is added to the database with rbind( ) and the updated database is saved to an RData file. These steps are repeated in the second section for carol_bp_db. The command add_bp_entry('bob', 121, 72) will add a new entry to the database bob_bp_db containing today's date, a systolic blood pressure value of 121, and a diastolic blood pressure value of 72.

It is important to understand that the database file is open and only available inside this function and not in your main R session. Even if the database is active in your current R session, this instance of the database will be opened, updated, and saved without altering the open version in your main R session.

The function view_bp_chart( ) creates a line chart of the systolic and diastolic blood pressure values in the selected blood pressure database. The function consists of three sections. The first section charts bob_bp_db. The second section charts carol_bp_db. The third section adds common details to either chart. The details of how the charts are defined and presented are worth reviewing and understanding.

view_bp_chart <- function(DB) { # chart the bp data
if(DB == 'bob') {
    load('bob_bp_db.RData') # open the database
    startDate <- bob_bp_db[1,1] # get the first date in the database to position the legend correctly
    # plot a line chart of the systolic bp [note that xaxt='n' supresses the x-axis values]
    plot(bob_bp_db$Systolic ~ bob_bp_db$Date, col='red', type = 'l', ylim = c(50,145), xaxt='n',

The function view_bp_chart( ) has one input argument that identifies which blood pressure database (bob or carol) to chart. The designated RData file is opened. The first date entry is saved so the chart legend can be properly positioned later in the function. A chart is plotted for the systolic blood pressure values with the command

plot(bob_bp_db$Systolic ~ bob_bp_db$Date, col='red', type = 'l', ylim = c(50,145), xaxt='n',

main = 'Bob\'s Blood Pressure', xlab = 'Date', ylab = 'BP')

Let's examine each argument of this plot( ) function call. The first argument [bob_bp_db$Systolic ~ bob_bp_db$Date] designates bob_bp_db$Date as the x-axis and bob_bp_db$Systolic as the y-axis of our chart. This argument uses the R model formula notation. The format for this notation is response ~ predictor. For the plot( ) function, the predictor defines the x-axis variable and the response defines the y-axis variable. The second argument [col='red'] sets the color of the plot to red. The third argument [type = 'l'] sets the plot type to line [the default type is dot]. The fourth argument [ylim = c(50,145)] sets the limit values for the y-axis. R normally sets the axis limits based on the maximum and minimum values of the data being plotted. If you wish specific maximum and minimum limits, as we do here, you must designate them using the ylim [or xlim] arguments. These limits will be maximum and minimum values in the ranges you want your chart to show. In this example we are designating specific minimum and maximum values for the y-axis so we can set the location for our chart legend later. The fifth argument [xaxt='n'] suppresses the default x-axis index values. We will add custom labels in another command and we must suppress the default axis values to prevent conflicting duplicate values. The sixth argument [main = 'Bob\'s Blood Pressure'] adds the chart title. Notice how the apostrophe is designated using the symbol set \'. The final two arguments [xlab = 'Date', ylab = 'BP'>] assign the labels for the x-axis and y-axis.

The all of the remaining commands use functions that add content to the current existing chart rather than creating a new chart. The axis( ) command adds a customized x-axis to our chart.

axis(1, bob_bp_db$Date, labels = format(bob_bp_db$Date, "%b-%d"), las = 2, cex.axis = .7)

The first argument in the axis( ) function identifies which side of the chart this function will customize. The possible argument values are: 1 = below, 2 = left, 3 = above, and 4 = right. The second argument [bob_bp_db$Date] identifies the data that will define the axis index values. The third argument [labels = format(bob_bp_db$Date, "%b-%d")] defines the format for the index value. This example uses a month-date format. The fourth argument [las = 2] defines the index label orientation in relation to the axis. The argument values identify whether the labels are parallel (=0) or perpendicular(=2) to axis. The final argument [cex.axis = .7] is the magnification of axis annotation relative to cex, where cex is the number indicating the amount by which plotting text and symbols should be scaled relative to the default. 1=default, 1.5 is 50% larger, 0.5 is 50% smaller, etc. In this example, the axis index labels are 70% of normal size.

The next command adds another line plot to our chart. This line will chart Bob's diastolic blood pressure.

lines(bob_bp_db$Diastolic ~ bob_bp_db$Date, col = 'green')

The lines( )uses two arguments in this example. The first argument [bob_bp_db$Diastolic ~ bob_bp_db$Date] uses the R model formula notation, bob_bp_db$Diastolic is identified as the response and bob_bp_db$Date is identified as the predictor. Notice that this function uses the same predictor variable as our plot( ) function above. This guarantees that this added line plot will coincide with the line plot in the main chart. The second argument [col = 'green'] sets the color of the new line to green. This completes the section of commands for Bob's individual chart. The next six commands will add generic content to either chart. This concludes the portion of the function creating the main blood pressure chart. This code is repeated if the other database is selected for charting.

The final block of code adds common features to the existing blood pressure chart. The function grid() adds a light grid to the background of the chart area. The function abline( ) is used four times. The first use inserts a horizontal dotted red line at the maximum normal systolic blood pressure value of 120. The second use inserts a heavy horizontal dotted red line at the borderline high systolic blood pressure value of 140. The third use inserts a horizontal dotted green line at the maximum normal diastolic blood pressure value of 80. The fourth use inserts a heavy horizontal dotted green line at the borderline high diastolic blood pressure value of 80. These reference lines enhance a viewer's understand what the chart values can mean in the context of normal and high blood pressure values.

The final command adds a legend to the chart so the viewer knows what the charted lines represent.

legend(as.Date(startDate),115, c('Systolic','Diastolic'),lty = c(1,1), col = c('red','green'))

The first two arguments of legend( ) [as.Date(startDate),115] identify the x-y location of the upper left corner of the legend box [as.Date(startDate),115]. These values are expressed in terms of the x and y axis values. In this example, the x coordinate is a date saved when the database was loaded at the beginning of the function view_bp_chart( )[the first date in the selected database]. The y coordinate is a blood pressure value. The value 115 was chosen so the legend is well placed vertically on the chart. The third argument [c('Systolic','Diastolic')] designates the text labels for the legend items. These labels are stored in a list created with the function c( ). The fourth argument [lty = c(1,1)] designates that the legend items should be shown as lines. The final argument [col = c('red','green')] designates the color for each legend item.

Here is what the chart looks like for the example database bob_bp_db.RData.

This little project is a good demonstration of several very useful R skills and techniques.

Return to the R Learning Infrastructure Home Web Page