Data Analysis and Visualisation Exercises Using R

The following exercises are aimed at aiding the learner use R for performing some types of data analysis and visualisation techniques that are illustrative of various approaches useful in most data projects.

For each exercise, the learner is expected to create either an R script or an R Markdown document to perform the tasks requested.

Exercise 1: Statistical methods for assessing agreement between two methods of clinical measurement

This exercise is based on:

Bland, J. M. & Altman, DouglasG. Statistical Methods For Assessing Agreement Between Two Methods Of Clinical Measurement. Lancet 327, 307–310 (1986).

The dataset used in the paper can be accessed from the teaching_datasets repository. The URL to the .dat file is https://raw.githubusercontent.com/OxfordIHTM/teaching_datasets/main/ba.dat.

The ba dataset contains peak expiratory flow rate (PEFR) measurements (in litres per minute) taken with a Wright peak flow metre (Wright) and a Mini-Wright peak flow metre (Mini-Wright). This is the same data that is presented in the referenced Lancet article above.

For this exercise, use the R script and/or R Markdown titled exercise1.

Tasks:

1. Read the ba.dat dataset from the GitHub URL provided above.

2. Create a function that will calculate the following metrics required by the Bland and Altman plot as described in the article referenced above:

   1. Mean of the per subject measurements made by the Wright and the Mini-Wright;

   2. Difference between the per subject measurements made by the Wright and the Mini-Wright;

   3. Mean of the difference between the per subject measurements made by the Wright and the Mini-Wright; and,

   4. Lower and upper limits of agreement between the per subject measurements made by the Wright and the Mini-Wright.

3. Develop a function that creates a Bland and Altman plot as described in the article.

4. Create a short report that shows

   1. A table of the mean and difference of per subject measurements made by the Wright and the Mini-Wright;

   2. A Bland and Altman plot; and,

   3. A short discussion of the findings with regard to the agreement between the Wright and Mini-Wright tools.

Exercise 2: Performing tests for correlations and associations between variables

This exercise uses the fem.dat dataset accessible from the teaching_datasets repository. The URL to the .dat file is https://raw.githubusercontent.com/OxfordIHTM/teaching_datasets/main/fem.dat. The fem.dat dataset is from 118 female pyschiatric patients.

The variables in the dataset are:

Variable Name	Variable Description
ID	Patient ID
AGE	Age in years
IQ	IQ score
ANX	Anxiety (1 = none, 2 = mild, 3 = moderate, 4 = severe)
DEP	Depression (1 = none, 2 = mild, 3 = moderate or severe)
SLP	Sleeping normally (1 = yes, 2 = no)
SEX	Lost interest in sex (1 = yes, 2 = no)
LIFE	Considered suicide (1 = yes, 2 = no)
WT	Weight change (kg) in previous 6 months

For this exercise, use the R script and/or R Markdown titled exercise2.

Tasks

1. Read the fem.dat dataset from the GitHub URL provided above;

2. Get a summary of the variables for AGE, IQ, and WT using tabular and graphical methods;

3. Test the correlation between weight gain (WT), age in years (AGE), and IQ (IQ) graphically and through formal statistical tests;

4. Calculate the odds ratio of someone with any form of anxiety considering suicide and calculate the odds ratio of someone with any for of depression considering suicide;

5. Specify a logistic regression model that looks into the correlation/association of considering suicide (LIFE) with depression (DEP), anxiety (ANX), abnormal sleep (SLP), and loss of interest in sex (SEX); and,

6. Create a short report that shows

   1. A tabular and graphical summary of the AGE, IQ, and WT variables;

   2. Results of correlation test between weight gain (WT), age in years (AGE), and IQ (IQ);

   3. Odds ratio of considering suicide and anxiety and odds ratio of considering suicde and depression;

   4. Results of logstic regression model for considering suicide with depression (DEP), anxiety (ANX), abnormal sleep (SLP), and loss of interest in sex (SEX); and,

   5. A short discussion of the findings.

Exercise 3: Time series plotting and analysis

This exercise uses the malaria.dat dataset accessible from the teaching_datasets repository. The URL to the .dat file is https://raw.githubusercontent.com/OxfordIHTM/teaching_datasets/main/malaria.dat. The malaria.dat dataset contains data on rainfall (in mm) and the number of cases of malaria reported from health centres in an administrative district of Ethiopia between July 1997 and July 1999

For this exercise, use the R script and/or R Markdown titled exercise4.

Tasks

1. Read the malaria.dat dataset from the GitHub URL given above;

2. Plot malaria cases over time and amount of rainfall over time in this administrative district of Ethiopia;

3. Decompose the malaria cases and amount of rainfall time series into its trend component, irregular component, and seasonal component;

4. Adjust the malaria cases and amount of rainfall time series for seasonality; and,

5. Create a short report that shows

   1. A plot of the malaria cases and amount of rainfall over time overload on top of the other;

   2. A table and plot of the trend component, seasonal component, and irregular/random component of the malaria cases and amount of rainfal time series;

   3. A table and plot of the seasonally adjusted time series for malaria cases and amount of rainfall; and,

   4. A short discussion of the findings.

Exercise 4: Performing text mining and topic modelling

This exercise uses the text data on COVID-related resolutions issued from 28 January 2020 to 30 July 2020 by the Inter-Agency Task Force for the Management of Emerging Infectious Diseases (IATF) in the Philippines.

The data is accessible through the covidphtext package available from GitHub.

To install this package, you will first need to install the remotes package:

install.packages("remotes")

and then install the covidphtext package from GitHub as follows:

remotes::install_github("como-ph/covidphtext")

Once installed, the full IATF resolutions dataset can be obtained by

library(covidphtext)
iatf_data <- combine_iatf()

For this exercise, use the R script and/or R Markdown titled exercise4.

Tasks

1. Load the IATF resolutions text data as describe above;

2. What is the general trend of sentiments expressed in the various resolutions over time?

3. What are the most common words (by sentiment) used in the various resolutions?

4. What general themes can be elicited from the various resolutions in the dataset? How do these themes relate to the general approach taken by the Philippine government nationally to manage the COVID-19 pandemic?

5. Create a short report that shows

   1. various relevant plots of sentiment trend over time;

   2. various relevant plots and tables showing term frequency;

   3. various relevant plots and tables showing topics modelled from the resolutions; and,

   4. A short discussion of the findings.