GEA1000 QUANTITATIVE REASONING WITH DATA
Tutorial 3
Please work on the problems before coming to class. In class, you will engage in group work.
1. For this question we will be working with the data set Singapore_Weather_Data.csv
taken from [Link]. The data set shows us the daily weather details for most days
in 2022 for 5 different parts of Singapore.
a) Using a suitable plot, compare the spread of the daily mean temperature between
the 5 stations. Based on the plot, which location would you stay at for the coolest
overall temperature?
b) Using summary statistics and a suitable plot, describe the symmetry or skewness
for the daily total rainfall at Sentosa Island.
c) Calculate the mean for the total daily rainfall and the daily mean temperature for
each of the 5 different stations, to 3 decimal places. Fill up the file
Ecological_Correlation.csv with your answers accordingly. Using this file, calculate
the ecological correlation between total daily rainfall and daily mean temperature
over the 5 different stations and describe its strength. With the ecological
correlation, what can we infer about the overall correlation between the two
variables at the individual level?
2. As part of an ongoing project to analyze global demographic trends, researchers have
compiled a simulated dataset, population_age.csv, generated by an intern. This
dataset focuses on key indicators such as median age and total population. While this
data serves as a valuable resource for applications like public policy planning,
economic strategies, and projections for future healthcare and educational needs, it
is important to note that the dataset was simulated for the purpose of this tutorial
and may contain inaccuracies or errors.
Median_Age serves as a pivotal measure, showing the age that divides the population
in two parts of equal size - there are as many persons in that region with ages above
the median as there are ages below1 the median. It is an indicator used to understand
societal aging, potential workforce demographics, and consumer behavior patterns.
Population quantifies the total number of individuals residing within each region,
influencing numerous aspects of national planning, including infrastructure
development, resource management, and environmental strategies.
1
People with age = Median_Age are not included in the top or bottom half.
However, preliminary reviews of the data set have revealed some inconsistencies,
particularly concerning median age values.
Objective: The primary goal of this analysis is to scrutinize the distribution of
Median_Age and Population, identifying logical inconsistencies and significant outliers
that could impact the validity of subsequent analyses. By addressing these data
integrity issues, the research aims to refine the data set, ensuring that further
demographic studies and policy recommendations are based on accurate and reliable
data.
Approach: The analysis will begin by charting the distributions of both Median_Age
and Population to visually identify and statistically confirm any aberrant or anomalous
values. This step is crucial in maintaining the integrity of demographic data, which
must be free from significant errors to support effective decision-making and accurate
forecasting.
a) Describe the distributions of Median_Age and Population in the data set using
i. suitable summary statistics (5-number summary, SD and IQR)
ii. visualisation tools (Histogram and Boxplot).
Are there any noticeable trends or outliers that could affect the analysis?
Hint: To make the histogram more readable, consider dividing the population
values by 1,000,000 to display the data in millions.
b) Before attempting this part, remove data points with Median_age below 152.
i. Construct a scatter plot of Population vs Median_Age. What do you
observe about the relationship between these two variables?
ii. Now construct another scatter plot, this time for ln(Population) vs
Median_Age.
(Note: ln() is the natural logarithm. You could also try using logarithm of
base 10, that is, log10(Population)). What do you notice about the
relationship shown in this scatter plot?
iii. Calculate and compare the correlation coefficients in parts b(i) and b(ii).
iv. Share a brief observation based on parts b(i), (ii) and (iii).
v. Discuss why taking the natural logarithm of the population might be
beneficial for this type of analysis. How does transforming the data affect
the interpretation of the results?
2
Countries do not typically have a median age below 15 years.
[Link]
c) Fit a linear regression model using Median_Age as the independent variable and
the natural log of Population as the dependent variable. What is the equation of
the fitted line?
d) Interpret the slope and intercept of the regression line you obtained for the
previous part. What do these coefficients/values imply about how Median_Age is
related to Population?
e) Using the regression model, predict the natural log of the Population for the given
Median_Age of 35 years. Convert this logarithmic prediction back into the actual
population estimate.