Assignment 07
Abdullah
2026-11-15
Setup and Data
library(ggplot2)
library(tidyr)
library(dplyr)
# Keep the CSV in the same folder as this .Rmd file
biomed <- read.csv("biomed_visualization_data.csv")
# Order categorical variables so legends/axes read logically
biomed$Diagnosis <- factor(biomed$Diagnosis, levels = c("Healthy", "Prediabetes", "Type1", "Type2"))
biomed$Gender <- factor(biomed$Gender, levels = c("F", "M"))
biomed$Treatment <- factor(biomed$Treatment, levels = c("Control", "DrugA"))
biomed$Cancer_Type <- factor(biomed$Cancer_Type, levels = c("Breast", "Colon", "Lung"))
str(biomed)## 'data.frame': 120 obs. of 12 variables:
## $ Patient_ID : chr "P001" "P002" "P003" "P004" ...
## $ BMI : num 28 25.4 28.6 32.1 25.1 25.1 32.3 29.1 24.1 28.2 ...
## $ BP_Sys : num 142 116 151 109 139 163 115 122 131 122 ...
## $ BP_Dia : num 74 81 87 91 70 79 77 75 100 86 ...
## $ Glucose : num 123 148 107 120 127 ...
## $ Cholesterol: num 206 177 154 217 184 ...
## $ Insulin : num 15.8 7.4 16.3 18.8 14.1 21.4 8.1 5.8 3.1 19.5 ...
## $ Gender : Factor w/ 2 levels "F","M": 1 2 1 1 2 2 2 2 1 1 ...
## $ Diagnosis : Factor w/ 4 levels "Healthy","Prediabetes",..: 3 4 2 1 3 3 3 2 1 4 ...
## $ Treatment : Factor w/ 2 levels "Control","DrugA": 2 2 1 1 2 2 2 1 1 1 ...
## $ Cancer_Type: Factor w/ 3 levels "Breast","Colon",..: 3 2 3 3 3 1 2 2 3 1 ...
## $ Stage : int 4 3 4 1 4 2 1 2 2 1 ...Note on axis convention: following Problem 1’s explicit wording (“BP_Sys (y-axis) vs BMI (x-axis”)), every “A vs B” plot below puts A on the y-axis and B on the x-axis. The dataset contains 120 patients.
Section A — Exploring Clinical and Metabolic Data
Problem 1: Blood Pressure vs BMI (colored by Gender)
ggplot(biomed, aes(x = BMI, y = BP_Sys, color = Gender)) +
geom_point(size = 2.5, alpha = 0.8) +
labs(
title = "Systolic Blood Pressure vs BMI by Gender",
x = "Body Mass Index (BMI, kg/m^2)",
y = "Systolic Blood Pressure (mmHg)"
) +
theme_minimal()
Answer/Interpretation: The points are widely scattered with Pearson r = 0.1, i.e., only a very weak positive BMI–SBP association in this cohort. Both genders overlap across the BMI range; a few high-SBP outliers (e.g., 188 mmHg) are visible.
Problem 2: Glucose vs Cholesterol (colored by Diagnosis)
ggplot(biomed, aes(x = Cholesterol, y = Glucose, color = Diagnosis)) +
geom_point(size = 2.5, alpha = 0.8) +
labs(
title = "Glucose vs Cholesterol by Diagnostic Group",
x = "Cholesterol (mg/dL)",
y = "Glucose (mg/dL)"
) +
theme_minimal()
Answer/Interpretation: Contrary to the suspected positive correlation, r = -0.01 — essentially no linear relationship between glucose and cholesterol here. The four diagnostic groups are intermixed rather than separated.
Problem 3: Distribution of BMI (30-bin histogram)
ggplot(biomed, aes(x = BMI)) +
geom_histogram(bins = 30, fill = "steelblue", color = "white") +
labs(
title = "Distribution of BMI in the Study Population",
x = "BMI (kg/m^2)",
y = "Number of Patients"
) +
theme_minimal()
Answer/Interpretation: BMI is roughly bell-shaped and centered near the mean of 25.7 kg/m^2 (SD = 3.7), broadly consistent with an approximately normal distribution, with mild right skew from a few high-BMI patients (up to 35.9).
Problem 4: Cholesterol Density by Gender
ggplot(biomed, aes(x = Cholesterol, color = Gender, fill = Gender)) +
geom_density(alpha = 0.35) +
labs(
title = "Cholesterol Distribution by Gender",
x = "Cholesterol (mg/dL)",
y = "Density"
) +
theme_minimal()
Answer/Interpretation: The two density curves overlap almost completely (mean cholesterol is ~205 mg/dL in females and ~205 mg/dL in males) — no meaningful gender difference in cholesterol in this cohort.
Problem 5: Glucose Across Diagnostic Groups (boxplot)
ggplot(biomed, aes(x = Diagnosis, y = Glucose, fill = Diagnosis)) +
geom_boxplot(alpha = 0.7) +
labs(
title = "Fasting Glucose Across Diagnostic Groups",
x = "Diagnosis",
y = "Glucose (mg/dL)"
) +
theme_minimal()
Answer/Interpretation: Prediabetes shows the highest median glucose and Healthy the lowest, but the boxes overlap substantially, so glucose alone does not cleanly separate the four groups in this simulated cohort.
Section B — Combining Layers and Adding Complexity
Problem 6: Glucose Boxplot + Jitter
ggplot(biomed, aes(x = Diagnosis, y = Glucose, fill = Diagnosis)) +
geom_boxplot(alpha = 0.7, outlier.shape = NA) + # hide outliers: jitter shows every point
geom_jitter(width = 0.2, alpha = 0.5) +
labs(
title = "Glucose by Diagnosis: Boxplot with Individual Patients",
x = "Diagnosis",
y = "Glucose (mg/dL)"
) +
theme_minimal()
Answer/Interpretation: Jittering reveals all 120 individual measurements, exposing the within-group spread (roughly 52–187 mg/dL) that the boxplot alone hides.
Problem 7: Systolic & Diastolic BP in One Violin Plot (long format)
bp_long <- biomed %>%
pivot_longer(cols = c(BP_Sys, BP_Dia),
names_to = "BP_Type", values_to = "Value")
head(bp_long) # rendered output: the reshaped 'long' data## # A tibble: 6 × 12
## Patient_ID BMI Glucose Cholesterol Insulin Gender Diagnosis Treatment
## <chr> <dbl> <dbl> <dbl> <dbl> <fct> <fct> <fct>
## 1 P001 28 123 206. 15.8 F Type1 DrugA
## 2 P001 28 123 206. 15.8 F Type1 DrugA
## 3 P002 25.4 148. 177. 7.4 M Type2 DrugA
## 4 P002 25.4 148. 177. 7.4 M Type2 DrugA
## 5 P003 28.6 107. 154. 16.3 F Prediabetes Control
## 6 P003 28.6 107. 154. 16.3 F Prediabetes Control
## # ℹ 4 more variables: Cancer_Type <fct>, Stage <int>, BP_Type <chr>,
## # Value <dbl>ggplot(bp_long, aes(x = BP_Type, y = Value, fill = Gender)) +
geom_violin(alpha = 0.7, position = position_dodge(0.8)) +
scale_x_discrete(labels = c(BP_Sys = "Systolic", BP_Dia = "Diastolic")) +
labs(
title = "Distribution of Systolic and Diastolic BP by Gender",
x = "Blood Pressure Type",
y = "Blood Pressure (mmHg)"
) +
theme_minimal()
Answer/Interpretation: Systolic values center around ~131 mmHg and diastolic around ~82 mmHg. Male and female violins have very similar shapes at both levels — no clear gender separation.
Problem 8: Patient Count per Diagnosis (bar chart)
ggplot(biomed, aes(x = Diagnosis)) +
geom_bar(fill = "skyblue", color = "black") +
geom_text(stat = "count", aes(label = after_stat(count)), vjust = -0.4) +
labs(
title = "Number of Patients per Diagnostic Group",
x = "Diagnosis",
y = "Count"
) +
theme_minimal()
## Diagnosis n
## 1 Healthy 29
## 2 Prediabetes 28
## 3 Type1 37
## 4 Type2 26Answer/Interpretation: Type1 = 37, Healthy = 29, Prediabetes = 28, Type2 = 26 patients (total n = 120). Groups are fairly balanced.
Problem 9: Cholesterol vs Glucose Faceted by Treatment
ggplot(biomed, aes(x = Glucose, y = Cholesterol, color = Diagnosis)) +
geom_point(size = 2.5, alpha = 0.8) +
facet_wrap(~Treatment) +
labs(
title = "Cholesterol vs Glucose by Diagnosis, Faceted by Treatment Arm",
x = "Glucose (mg/dL)",
y = "Cholesterol (mg/dL)"
) +
theme_minimal()
Answer/Interpretation: Control (n = 65) and DrugA (n = 55) panels show the same diffuse cloud — no visible treatment-arm difference in the glucose–cholesterol relationship.
Problem 10: Publication Palette (Set2) + theme_classic()
ggplot(biomed, aes(x = Glucose, y = Cholesterol, color = Diagnosis)) +
geom_point(size = 2.5, alpha = 0.9) +
facet_wrap(~Treatment) +
scale_color_brewer(palette = "Set2") +
labs(
title = "Cholesterol vs Glucose by Diagnosis and Treatment",
x = "Glucose (mg/dL)",
y = "Cholesterol (mg/dL)",
color = "Diagnosis"
) +
theme_classic()
Answer/Interpretation: Same figure as Problem 9, now
with the colorblind-friendly ColorBrewer Set2 palette
and the clean, axis-only theme_classic() look suitable for
publication.
Section C — Clinical Summaries and Patterns
Problem 11: Insulin by Diagnosis (boxplot + jitter)
ggplot(biomed, aes(x = Diagnosis, y = Insulin, fill = Diagnosis)) +
geom_boxplot(alpha = 0.7, outlier.shape = NA) +
geom_jitter(width = 0.2, alpha = 0.5) +
labs(
title = "Serum Insulin Levels Across Diagnostic Groups",
x = "Diagnosis",
y = "Insulin (uU/mL)"
) +
theme_minimal()
Answer/Interpretation: Insulin distributions overlap heavily across groups (Type2 has the highest mean, Type1 the lowest), so the expected insulin-resistance pattern is weak in this simulated data. Data-quality note: two records (P047 = −1.5, P069 = −1.3) have negative insulin values, which is physiologically impossible — an artifact of the simulation. They are kept as-is because the assignment does not ask for cleaning, but they should be flagged/excluded in a real analysis.
Problem 12: Systolic & Diastolic BP by Gender
bp_long <- biomed %>%
pivot_longer(cols = c(BP_Sys, BP_Dia),
names_to = "BP_Type", values_to = "Value")
ggplot(bp_long, aes(x = Gender, y = Value, fill = BP_Type)) +
geom_boxplot(position = position_dodge(0.8), alpha = 0.8) +
scale_fill_manual(values = c(BP_Sys = "#E69F00", BP_Dia = "#56B4E9"),
labels = c(BP_Sys = "Systolic", BP_Dia = "Diastolic")) +
labs(
title = "Comparison of Systolic and Diastolic BP by Gender",
x = "Gender",
y = "Blood Pressure (mmHg)",
fill = "BP Type"
) +
theme_minimal()
bp_long %>%
group_by(Gender, BP_Type) %>%
summarise(Mean_BP = round(mean(Value), 1), .groups = "drop") # rendered numeric answer## # A tibble: 4 × 3
## Gender BP_Type Mean_BP
## <fct> <chr> <dbl>
## 1 F BP_Dia 83.3
## 2 F BP_Sys 131.
## 3 M BP_Dia 81
## 4 M BP_Sys 131.Answer/Interpretation: The hypothesis that men have higher BP is not supported here: mean systolic is essentially equal (F ≈ 131.3 vs M ≈ 130.8 mmHg) and females actually have slightly higher mean diastolic BP (83.3 vs 81.0 mmHg).
Problem 13: Glucose vs Insulin by Cancer Type, Sized by Stage
ggplot(biomed, aes(x = Insulin, y = Glucose, color = Cancer_Type)) +
geom_point(aes(size = Stage), alpha = 0.8) +
geom_smooth(method = "lm", se = FALSE) + # one regression line per cancer type
labs(
title = "Glucose vs Insulin by Cancer Type (Point Size = Tumor Stage)",
x = "Insulin (uU/mL)",
y = "Glucose (mg/dL)",
color = "Cancer Type",
size = "Stage"
) +
theme_minimal()
Answer/Interpretation: The fitted trends differ by cancer type: the Breast group shows a negative glucose–insulin slope (r ≈ −0.31), while Colon and Lung show near-flat/slightly positive slopes — no consistent stage-wide biomarker correlation.
Section D — Integrative & Publication-Ready Visuals
Problem 14: Cardiometabolic Risk by Treatment Arm
ggplot(biomed, aes(x = Glucose, y = BMI, color = Diagnosis)) +
geom_point(size = 2.5, alpha = 0.8) +
facet_wrap(~Treatment) +
labs(
title = "BMI vs Glucose Across Treatment Arms",
x = "Glucose (mg/dL)",
y = "BMI (kg/m^2)",
color = "Metabolic Status"
) +
theme_minimal(base_size = 12) +
theme(
legend.position = "bottom",
plot.title = element_text(face = "bold"),
strip.text = element_text(face = "bold")
)
Answer/Interpretation: BMI and glucose are essentially unrelated (r = -0.04) in both arms, and diagnostic groups overlap within each facet — no treatment-specific cardiometabolic pattern emerges. The legend is titled “Metabolic Status” as required.
Problem 15: Publication-Ready Insulin Summary Figure
ggplot(biomed, aes(x = Diagnosis, y = Insulin, fill = Diagnosis)) +
geom_boxplot(alpha = 0.9, outlier.shape = NA) +
geom_jitter(width = 0.2, alpha = 0.5, color = "grey25") +
scale_fill_brewer(palette = "Set3") +
labs(
title = "Serum Insulin Distribution Across Diagnostic Groups",
subtitle = "Study Population: n = 120",
caption = "Source: Simulated Biomedical Dataset",
x = "Diagnosis",
y = "Serum Insulin (uU/mL)"
) +
theme_minimal() +
theme(
plot.title = element_text(hjust = 0.5, face = "bold"),
plot.subtitle = element_text(hjust = 0.5),
plot.caption = element_text(hjust = 0.5)
)
Answer/Interpretation: Final journal-style figure:
boxplot + jitter, RColorBrewer Set3 fill palette,
required title/subtitle/caption, theme_minimal(), and all
title text centered with hjust = 0.5. Insulin distributions
are broadly similar across the four diagnostic groups.
Reproducibility
## R version 4.4.3 (2025-02-28)
## Platform: x86_64-conda-linux-gnu
## Running under: Ubuntu 24.04.5 LTS
##
## Matrix products: default
## BLAS/LAPACK: /home/neolith/anaconda3/envs/rstudio_env/lib/libopenblasp-r0.3.33.so; LAPACK version 3.12.0
##
## locale:
## [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
## [3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8
## [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
## [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
## [9] LC_ADDRESS=C LC_TELEPHONE=C
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
##
## time zone: Asia/Dhaka
## tzcode source: system (glibc)
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] dplyr_1.2.1 tidyr_1.3.2 ggplot2_4.0.4
##
## loaded via a namespace (and not attached):
## [1] vctrs_0.7.3 cli_3.6.6 knitr_1.52 rlang_1.3.0
## [5] xfun_0.59 purrr_1.2.2 generics_0.1.4 S7_0.2.2
## [9] jsonlite_2.0.0 labeling_0.4.3 glue_1.8.1 htmltools_0.5.9
## [13] sass_0.4.10 scales_1.4.0 rmarkdown_2.32 grid_4.4.3
## [17] tibble_3.3.1 evaluate_1.0.5 jquerylib_0.1.4 fastmap_1.2.0
## [21] yaml_2.3.12 lifecycle_1.0.5 compiler_4.4.3 RColorBrewer_1.1-3
## [25] pkgconfig_2.0.3 rstudioapi_0.19.0 farver_2.1.2 digest_0.6.39
## [29] R6_2.6.1 utf8_1.2.6 tidyselect_1.2.1 pillar_1.11.1
## [33] magrittr_2.0.5 bslib_0.11.0 withr_3.0.3 tools_4.4.3
## [37] gtable_0.3.6 cachem_1.1.0