Lab 3: Proportions and Hierarchies

STAT 80B: Data Visualization - Edible Plants Edition

Author

Overview

Due: Thursday, at the end of class
Points: 10

Submission: PDF with visualizations + written critique

In this lab, you’ll practice creating and critiquing visualizations of proportions and hierarchical data using the TidyTuesday Edible Plants Database. You’ll create multiple visualizations of the same data and then analyze which ones work best and why.

Learning Objectives

By completing this lab, you will:

Create different types of proportion visualizations (pie charts, stacked bars, side-by-side bars)
Build hierarchical visualizations (treemaps, sunburst charts)
Critically evaluate the strengths and weaknesses of different visualization choices
Develop design judgment for choosing appropriate visualizations
Explore patterns in edible plant diversity across families, regions, and uses

Dataset: Edible Plants Database

For this lab, you’ll use the TidyTuesday Edible Plants Database (February 3, 2026, Week 5).

About the Data

This dataset contains information about edible plants from around the world, including:

Plant families: Taxonomic classification (e.g., Fabaceae, Solanaceae, Rosaceae)
Plant cultivation: Cultivation class. Brassica, Legume, etc.
Sun light: How much sunlight the plant requires.
Common and scientific names: Plant identification
Possibly: nutritional information, culinary uses, growing conditions

Loading the Data

# Install tidytuesdayR if needed
# install.packages("tidytuesdayR")

# Load required packages
library(tidyverse)
library(treemapify)  # For treemaps

# Load data directly from TidyTuesday
tuesdata <- tidytuesdayR::tt_load('2026-02-03')
# OR
tuesdata <- tidytuesdayR::tt_load(2026, week = 5)

edible_plants <- tuesdata$edible_plants

# View structure
glimpse(edible_plants)
head(edible_plants)
names(edible_plants)  # See all column names

# View the readme for more information about the dataset
readme(tuesdata)

# Import required libraries
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import squarify  # For treemaps: pip install squarify
import plotly.express as px  # For interactive plots

# Set style
sns.set_style("whitegrid")
plt.rcParams['figure.dpi'] = 300

# Load data
edible_plants = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2026/2026-02-03/edible_plants.csv')

# View structure
print(edible_plants.head())
print(edible_plants.info())
print(edible_plants.columns)
print(edible_plants.describe())

Open Tableau Desktop
Download CSV from GitHub https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2026/2026-02-03/edible_plants.csv
- In Tableau: Connect to Data → Text file → select downloaded CSV

Another option - Use R to download:

Run R code from the R section
Export: write_csv(edible_plants, "edible_plants.csv")
Load into Tableau

Exploring the Data

First step: Explore what’s actually in your dataset!

# In R - explore the structure
glimpse(edible_plants)
summary(edible_plants)

# Check unique values for key columns
edible_plants %>% count(family) %>% arrange(desc(n))
edible_plants %>% count(plant_type) %>% arrange(desc(n))  # or whatever the column is called

In every tool: explore all the variables, how many of them are categorical, how many are numerical, what do they represent, and which questions could you answer using them?

Part 1: Visualizing Plant Diversity Proportions (4 points)

Task

Create FOUR different visualizations showing compositional patterns in the edible plants data.

Choose ONE of these approaches:

Option A: Plant Family Distribution

Analyze which plant families contribute the most edible species.

Create these four visualizations:

Pie chart - showing proportion of plants by family (top 8-10 families)
Bar chart - ordered by count (most common families first)
Bar chart - different ordering (alphabetical, or grouped by botanical order)
Donut chart - alternative circular visualization with top families highlighted

Example questions to explore:

What proportion of edible plants come from the Fabaceae (legume) family?
How dominant is the Rosaceae family (roses, apples, cherries)?
Are edible plants concentrated in a few families or widely distributed?

Option B: Plant Type Comparison Across Water requirement

Compare the distribution of plant types across different categories of how much water the plant requires.

Create these four visualizations:

Stacked bar chart - comparing plant types (vegetables, fruits, grains, herbs) by water requirement
100% stacked bar chart - proportional distribution showing composition differences
Side-by-side bar chart - for direct comparison of each type across categories
Small multiples - separate pie or bar charts for each categories

Example questions to explore:

Does fruit species need more water?
Which water requirement level have the greatest diversity of edible plants?
How does plant type composition vary by water requirements?

Requirements

Each visualization must:

✅ Be properly labeled (title, axis labels, legend if needed)
✅ Show counts or percentages clearly
✅ Use appropriate colors (consider green palette for plants, or color by category)
✅ Be readable (appropriate size, font, spacing)
✅ Include a brief caption (2-3 sentences) explaining what it shows and one key insight

Data Preparation Tips

You’ll need to:

Count occurrences: Use count() in R or value_counts() in Python
Calculate percentages: count / total * 100
Filter to top categories: Too many small slices make pie charts unreadable
- Consider showing only top 10-12 families/categories
- Group remaining into “Other”
Handle missing data: Decide how to deal with NAs
Create clear labels: Shorten long scientific names if needed

Deliverable

4 visualizations (can be arranged using Dashboard on Tableau)
Each with a 2-3 sentence caption that includes one specific insight
Professional appearance with consistent styling

Part 2: Visualizing Plant Classification Hierarchies (4 points)

Task

Create ONE hierarchical visualization showing the nested structure of edible plant classifications.

Recommended Hierarchy Structures

Option 1: Taxonomic Hierarchy

Level 1: Taxonomy (e.g., Fabaceae, Rosaceae, Solanaceae)
Level 2: Individual species count OR specific edible parts

Best for: Understanding botanical diversity and which families dominate our food system

Option 2: Two-Level Simplified

Level 1: Taxonomy (top 15-20 families)
Level 2: Water needs within each family

Best for: Clear, focused visualization when you have limited space

Visualization Type Options

Choose ONE:

Treemap (recommended) - rectangles sized by count, excellent for comparing categories
Sunburst chart - circular rings showing hierarchy, good for part-whole relationships
Icicle plot - vertical rectangles, like a treemap rotated
Circle packing - nested circles, aesthetically pleasing but harder to compare sizes

Requirements

Your visualization must:

✅ Show at least 2 levels of hierarchy clearly
✅ Use size to encode importance (larger = more species, more diversity)
✅ Use color strategically:
- Color by plant family (with consistent palette)
- OR color by plant type (vegetables=green, fruits=red/orange, etc.)
- OR color by region
- NOT random colors!
✅ Label major categories clearly (family names, major groups)
✅ Include a clear, descriptive title
✅ Have a 2-3 sentence caption explaining the main pattern or insight

Deliverable

1 hierarchical visualization
Professional appearance with thoughtful design
Caption explaining what patterns are visible

Part 3: Critical Evaluation (2 points)

Task

Write a 2-page critique (approximately 200-300 words) analyzing your visualizations.

Address the following questions:

Section 1: Comparing Proportion Visualizations

For your four proportion visualizations:

Which visualization is most effective for showing water requirement diversity patterns?
Which visualization is least effective?
For what audience or purpose would each be appropriate?
What design choices did you make and why?

Section 2: Evaluating Your Hierarchical Visualization

For your hierarchical visualization:

What story does it tell about edible plant diversity?
What works well in your design?
What could be improved?
Would a different visualization type be better?

Deliverable

Written analysis
Thoughtful, specific critique with concrete examples from your visualizations
References to visualization principles from class (pre-attentive attributes, color theory, etc.)
Professional writing (well-organized, proofread!)

Grading Rubric

Part 1: Proportions (4 points)

Criterion	Points	What We’re Looking For
Four distinct visualizations	1.6	All four types completed; appropriate for botanical/count data; different enough to compare
Design quality	1.2	Professional appearance; thoughtful color choices; proper sizing; clean layout
Labels and annotations	0.8	Clear titles, axis labels, legends; counts/percentages shown appropriately; readable text
Captions with insights	0.4	2-3 sentence captions that explain the visualization AND state one specific insight about plants

Part 2: Hierarchy (4 points)

Criterion	Points	What We’re Looking For
Appropriate visualization choice	1	Chosen type suits the plant data and classification hierarchy
Hierarchy clearly shown	1	Multiple levels visible; parent-child relationships clear; well-organized structure
Design quality	1	Effective use of size (species count), color (families/types), layout; professional look
Labels and annotations	1	Major categories labeled; clear title; insightful 2-3 sentence caption about diversity patterns

Part 3: Critique (2 points)

Criterion	Points	What We’re Looking For
Comparison of proportion visualizations	0.8	Thoughtful comparison; specific examples; considers different audiences; identifies trade-offs
Evaluation of hierarchical visualization	0.8	Honest self-assessment; identifies both strengths and weaknesses; considers alternatives
Writing quality	0.4	Clear, well-organized; proper grammar; professional tone; references class concepts

Total: 10 points

Technical Instructions & Code Examples

R Code Examples

Data Preparation & Exploration

library(tidyverse)
library(treemapify)

# Load data
tuesdata <- tidytuesdayR::tt_load('2026-02-03')
edible_plants <- tuesdata$edible_plants

# FIRST: Explore your data!
glimpse(edible_plants)
head(edible_plants, 20)
names(edible_plants)

# Look at unique values for key columns
# (Adjust column names based on what you actually have!)
edible_plants %>% count(family, sort = TRUE)
edible_plants %>% count(plant_type, sort = TRUE)

# Option 1: Count plants by family
family_counts <- edible_plants %>%
  count(family, sort = TRUE) %>%
  mutate(percentage = n / sum(n) * 100)

# Filter to top 10 families for cleaner visualizations
top_families <- family_counts %>%
  slice_max(n, n = 10)

# Add "Other" category for remaining families
top_families_with_other <- family_counts %>%
  mutate(family_group = if_else(family %in% top_families$family, family, "Other")) %>%
  group_by(family_group) %>%
  summarize(n = sum(n)) %>%
  mutate(percentage = n / sum(n) * 100)

# Option 2: Count by edible part
edible_part_counts <- edible_plants %>%
  count(edible_part, sort = TRUE) %>%  # Adjust column name!
  mutate(percentage = n / sum(n) * 100)

# Option 3: Cross-tabulation - families by plant type
family_type <- edible_plants %>%
  count(family, plant_type) %>%
  group_by(family) %>%
  mutate(total = sum(n)) %>%
  ungroup() %>%
  filter(total >= 5)  # Only families with 5+ species

# Option 4: Regional comparison (if you have region data)
regional_comparison <- edible_plants %>%
  count(region, plant_type) %>%
  group_by(region) %>%
  mutate(percentage = n / sum(n) * 100)

Visualization 1: Pie Chart

# Basic pie chart - top plant families
ggplot(top_families, aes(x = "", y = n, fill = family)) +
  geom_bar(stat = "identity", width = 1) +
  coord_polar("y", start = 0) +
  theme_void() +
  labs(title = "Distribution of Edible Plants by Family (Top 10)",
       subtitle = "Based on TidyTuesday Edible Plants Database",
       fill = "Plant Family") +
  scale_fill_brewer(palette = "Set3") +  # or "Greens", "YlGn"
  geom_text(aes(label = paste0(round(percentage, 1), "%")),
            position = position_stack(vjust = 0.5),
            size = 3,
            color = "black")

ggsave("pie_chart_families.png", width = 8, height = 6, dpi = 300)

# Alternative: Pie chart with "Other" category
ggplot(top_families_with_other, aes(x = "", y = n, fill = family_group)) +
  geom_bar(stat = "identity", width = 1) +
  coord_polar("y", start = 0) +
  theme_void() +
  labs(title = "Edible Plant Families",
       fill = "Family") +
  scale_fill_viridis_d(option = "G", begin = 0.2, end = 0.9)

ggsave("pie_chart_with_other.png", width = 8, height = 6, dpi = 300)

Visualization 2: Bar Chart (Horizontal)

# Horizontal bar chart - easier to read family names
ggplot(top_families, aes(x = n, y = reorder(family, n), fill = family)) +
  geom_col(show.legend = FALSE) +
  labs(title = "Top 10 Edible Plant Families by Species Count",
       subtitle = "Number of edible species per plant family",
       x = "Number of Species",
       y = "Plant Family") +
  theme_minimal() +
  scale_fill_brewer(palette = "Set3") +
  geom_text(aes(label = paste0(n, " (", round(percentage, 1), "%)")),
            hjust = -0.1, size = 3.5) +
  scale_x_continuous(expand = expansion(mult = c(0, 0.15)))  # Make room for labels

ggsave("bar_chart_horizontal.png", width = 10, height = 6, dpi = 300)

Visualization 3: Bar Chart (Vertical, Different Order)

# Vertical bar chart - alphabetical order
ggplot(top_families, aes(x = reorder(family, family), y = n, fill = family)) +
  geom_col(show.legend = FALSE) +
  labs(title = "Top Plant Families - Alphabetical Order",
       x = "Plant Family",
       y = "Number of Edible Species") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 10)) +
  scale_fill_viridis_d(option = "viridis") +
  geom_text(aes(label = n), vjust = -0.5, size = 3)

ggsave("bar_chart_alphabetical.png", width = 10, height = 6, dpi = 300)

Visualization 4: Stacked or Grouped Bars

# If comparing across regions or types:
# Stacked bar chart
ggplot(regional_comparison, aes(x = region, y = n, fill = plant_type)) +
  geom_bar(stat = "identity", position = "stack") +
  labs(title = "Edible Plant Types by Region",
       x = "Region",
       y = "Number of Species",
       fill = "Plant Type") +
  theme_minimal() +
  scale_fill_brewer(palette = "Set2") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

ggsave("stacked_bar_regional.png", width = 10, height = 6, dpi = 300)

# 100% stacked bar
ggplot(regional_comparison, aes(x = region, y = percentage, fill = plant_type)) +
  geom_bar(stat = "identity", position = "fill") +
  labs(title = "Composition of Edible Plant Types by Region",
       x = "Region",
       y = "Percentage",
       fill = "Plant Type") +
  scale_y_continuous(labels = scales::percent_format(scale = 1)) +
  theme_minimal() +
  scale_fill_brewer(palette = "Set2")

ggsave("stacked_bar_percent.png", width = 10, height = 6, dpi = 300)

# Grouped (side-by-side) bars
ggplot(regional_comparison, aes(x = region, y = n, fill = plant_type)) +
  geom_bar(stat = "identity", position = "dodge") +
  labs(title = "Edible Plant Types by Region - Direct Comparison",
       x = "Region",
       y = "Number of Species",
       fill = "Plant Type") +
  theme_minimal() +
  scale_fill_brewer(palette = "Set2") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

ggsave("grouped_bar.png", width = 12, height = 6, dpi = 300)

Hierarchical Visualization: Treemap

# Treemap: Family > Plant Type > Count
# Prepare hierarchical data
treemap_data <- edible_plants %>%
  count(family, plant_type) %>%
  filter(n >= 2)  # Only include family-type combos with 2+ species

# Basic treemap
ggplot(treemap_data, 
       aes(area = n, fill = family, label = plant_type, subgroup = family)) +
  geom_treemap() +
  geom_treemap_text(colour = "white", 
                    place = "centre", 
                    grow = FALSE,
                    reflow = TRUE,
                    size = 10) +
  geom_treemap_subgroup_border(colour = "white", size = 3) +
  geom_treemap_subgroup_text(place = "centre", 
                             alpha = 0.7,
                             colour = "white",
                             fontface = "bold",
                             grow = TRUE,
                             size = 14) +
  labs(title = "Edible Plant Diversity: Family → Type",
       subtitle = "Size represents number of edible species",
       fill = "Plant Family") +
  theme(legend.position = "bottom",
        legend.text = element_text(size = 8)) +
  scale_fill_viridis_d(option = "G", begin = 0.1, end = 0.9) +
  guides(fill = guide_legend(nrow = 2))

ggsave("treemap_family_type.png", width = 14, height = 10, dpi = 300)

# Alternative: Just families (simpler)
family_counts_filtered <- family_counts %>%
  filter(n >= 3)  # Only families with 3+ species

ggplot(family_counts_filtered, 
       aes(area = n, fill = n, label = paste0(family, "\n", n))) +
  geom_treemap() +
  geom_treemap_text(colour = "white", 
                    place = "centre",
                    size = 12,
                    grow = TRUE) +
  labs(title = "Edible Plant Families by Species Count",
       subtitle = "Each rectangle represents a plant family, sized by number of edible species") +
  scale_fill_gradient(low = "#d4f1d4", high = "#0d5c0d", 
                      name = "Species\nCount") +
  theme(legend.position = "right")

ggsave("treemap_families_only.png", width = 12, height = 8, dpi = 300)

Creating Multi-Panel PDFs

# Save all plots to one PDF using gridExtra
library(gridExtra)

p1 <- # your pie chart ggplot object
p2 <- # your first bar chart
p3 <- # your second bar chart  
p4 <- # your stacked/grouped bar chart

# Arrange in grid
grid.arrange(p1, p2, p3, p4, ncol = 2,
             top = "Lab 3: Edible Plants Proportion Visualizations")

ggsave("all_proportion_viz.pdf", width = 14, height = 10)

# Alternative: patchwork package
library(patchwork)
(p1 | p2) / (p3 | p4) + 
  plot_annotation(title = 'Edible Plants Visualizations',
                  theme = theme(plot.title = element_text(size = 16, face = "bold")))

ggsave("all_proportion_viz_patchwork.pdf", width = 14, height = 10)

Python Code Examples

Data Preparation

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import squarify  # pip install squarify
import plotly.express as px

# Set style
sns.set_style("whitegrid")
plt.rcParams['figure.dpi'] = 300

# Load data
edible_plants = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2026/2026-02-03/edible_plants.csv')

# Explore
print(edible_plants.head())
print(edible_plants.info())
print(edible_plants.columns)

# Count by family
family_counts = edible_plants['family'].value_counts().reset_index()
family_counts.columns = ['family', 'count']
family_counts['percentage'] = (family_counts['count'] / family_counts['count'].sum()) * 100

# Top 10 families
top_families = family_counts.head(10)

# Count by edible part
edible_part_counts = edible_plants['edible_part'].value_counts().reset_index()
edible_part_counts.columns = ['edible_part', 'count']

# Cross-tabulation
family_type = pd.crosstab(edible_plants['family'], edible_plants['plant_type'])

Pie Chart

# Pie chart
plt.figure(figsize=(10, 8))
colors = sns.color_palette('Set3', len(top_families))

plt.pie(top_families['count'], 
        labels=top_families['family'],
        autopct='%1.1f%%',
        startangle=90,
        colors=colors,
        textprops={'fontsize': 11})

plt.title('Distribution of Edible Plants by Family (Top 10)', 
          fontsize=16, fontweight='bold', pad=20)
plt.axis('equal')
plt.tight_layout()
plt.savefig('pie_chart.png', bbox_inches='tight', dpi=300)
plt.show()

Horizontal Bar Chart

# Horizontal bar chart
plt.figure(figsize=(10, 8))
colors_bar = sns.color_palette('Set3', len(top_families))

plt.barh(top_families['family'], top_families['count'], color=colors_bar)
plt.xlabel('Number of Species', fontsize=12)
plt.ylabel('Plant Family', fontsize=12)
plt.title('Top 10 Edible Plant Families by Species Count', 
          fontsize=14, fontweight='bold')

# Add count labels
for i, (count, pct) in enumerate(zip(top_families['count'], top_families['percentage'])):
    plt.text(count + max(top_families['count'])*0.02, i, 
             f'{count} ({pct:.1f}%)', 
             va='center', fontsize=10)

plt.gca().invert_yaxis()  # Highest at top
plt.tight_layout()
plt.savefig('bar_chart_horizontal.png', dpi=300)
plt.show()

Stacked Bar Chart

# Stacked bar chart (example with plant types by region)
# First create a pivot table
pivot_data = edible_plants.groupby(['region', 'plant_type']).size().unstack(fill_value=0)

# Plot
ax = pivot_data.plot(kind='bar', stacked=True, 
                     figsize=(12, 6),
                     colormap='Set2')
plt.xlabel('Region', fontsize=12)
plt.ylabel('Number of Species', fontsize=12)
plt.title('Edible Plant Types by Region', fontsize=14, fontweight='bold')
plt.legend(title='Plant Type', bbox_to_anchor=(1.05, 1), loc='upper left')
plt.xticks(rotation=45, ha='right')
plt.tight_layout()
plt.savefig('stacked_bar.png', dpi=300)
plt.show()

# 100% stacked
pivot_pct = pivot_data.div(pivot_data.sum(axis=1), axis=0) * 100
ax = pivot_pct.plot(kind='bar', stacked=True,
                    figsize=(12, 6),
                    colormap='Set2')
plt.ylabel('Percentage', fontsize=12)
plt.title('Composition of Plant Types by Region (%)', fontsize=14, fontweight='bold')
plt.legend(title='Plant Type', bbox_to_anchor=(1.05, 1))
plt.tight_layout()
plt.savefig('stacked_bar_percent.png', dpi=300)
plt.show()

Treemap

# Static treemap using squarify
plt.figure(figsize=(14, 10))

# Prepare data
sizes = top_families['count'].tolist()
labels = [f"{row['family']}\n{row['count']} species" 
          for _, row in top_families.iterrows()]
colors_tree = sns.color_palette('Greens', len(top_families))

squarify.plot(sizes=sizes, 
              label=labels, 
              color=colors_tree,
              alpha=0.8,
              text_kwargs={'fontsize': 11, 'weight': 'bold', 'color': 'white'})

plt.title('Edible Plant Families by Species Count', 
          fontsize=16, fontweight='bold')
plt.axis('off')
plt.tight_layout()
plt.savefig('treemap.png', dpi=300)
plt.show()

# Interactive treemap using Plotly
# Create hierarchical data
tree_data = edible_plants.groupby(['family', 'plant_type']).size().reset_index(name='count')
tree_data = tree_data[tree_data['count'] >= 2]  # Filter small groups

fig = px.treemap(tree_data,
                 path=['family', 'plant_type'],
                 values='count',
                 title='Edible Plant Diversity: Family → Type',
                 color='family',
                 color_discrete_sequence=px.colors.qualitative.Set3)

fig.update_layout(font=dict(size=12))
fig.write_html('treemap_interactive.html')
fig.show()

# Save screenshot for PDF
fig.write_image('treemap_interactive.png', width=1400, height=1000)

Sunburst Chart (Interactive)

# Sunburst chart using Plotly
fig = px.sunburst(tree_data,
                  path=['family', 'plant_type'],
                  values='count',
                  title='Edible Plants: Hierarchical View',
                  color='family',
                  color_discrete_sequence=px.colors.qualitative.Set3)

fig.update_layout(font=dict(size=12))
fig.write_html('sunburst_interactive.html')
fig.show()

Tableau Instructions

For Proportion Visualizations

Pie Chart:

Drag Family to Color (in Marks card)
Drag Family to Angle (or use automatic measure)
Change mark type to Pie
Right-click on the worksheet → Duplicate to count: use COUNT([Family])
Add labels: Drag Family to Label, add percentage
Edit colors: Click Color → Edit Colors → choose palette

Bar Chart:

Drag Family to Rows
Drag CNT(Record) or count measure to Columns
Click Sort descending button
Drag Family to Color (optional)
Right-click axis → Add Reference Line if needed

Stacked Bar:

Drag Region to Columns
Drag Plant Type to Color
Drag count measure to Rows
Analysis → Stack Marks → On
For percentages: Right-click axis → Quick Table Calculation → Percent of Total

For Treemap

Create hierarchy: Right-click Family → Hierarchy → Create
Add Plant Type to the hierarchy
New worksheet: Change mark type to Square
Drag hierarchy to Detail
Drag count measure to Size
Drag Family to Color
Drag labels to Label
Format: Color → Edit Colors; Borders for clarity

Submission Instructions

What to Submit

Submit ONE PDF file containing:

Part 1: Four proportion visualizations (1-2 pages)
- All four visualizations clearly visible
- Each with a 2-3 sentence caption that includes one specific insight
Part 2: One hierarchical visualization (1 page)
- Your treemap/sunburst/etc.
- With 2-3 sentence caption about diversity patterns
Part 3: Written critique (1 pages)
- Well-organized with clear sections
- Professional writing

Total: 3-4 pages

Formatting Requirements

✅ PDF format only
✅ Filename: Lab3_YourLastName.pdf
✅ Readable visualizations: Not too small! Use high DPI (300+)
✅ Professional appearance: Consistent fonts, clean layout
✅ Pages numbered
✅ Your name on first page

How to Submit

Create all visualizations and save as high-quality images (PNG, 300 DPI)
Write your critique in Word/Google Docs
Combine everything into one PDF:
- Word: File → Save As → PDF
- Google Docs: File → Download → PDF
- Or use online PDF merger
Upload to Canvas before Thursday class ends

Tips for Success

General Tips

Start early! Explore the data first before committing to specific visualizations
Check column names - they may differ from examples, adapt code accordingly
Ask questions in office hours or on the discussion board
Be honest in your critique - self-reflection earns points!

Data Exploration Tips

Look at the readme - it has important context about the data
Count things first - understand distributions before visualizing
Filter strategically - show top 10-15 categories, combine rare ones into “Other”
Handle NAs - decide if you want to show missing data or exclude it

Visualization Tips

Less is more: Don’t try to show everything at once
Color matters: Use green for plants, or color by category consistently
Label clearly: Plant family names can be long - consider abbreviating or using horizontal bars
Tell a story: What’s the main message each visualization should convey?
Be consistent: Use same color scheme across your four proportion visualizations

Writing Tips

Be specific: “The Fabaceae family dominates with 47 species (23%)” is better than “One family is very common”
Reference concepts from class: Pre-attentive attributes, color theory, gestalt principles
Compare thoughtfully: Discuss trade-offs between visualization types
Connect to real world: What do these patterns mean for biodiversity, food security, agriculture?
Proofread! Typos hurt your credibility

Frequently Asked Questions

Q: What if column names are different from the examples?
A: Expected! Use glimpse(), head(), or info() to see actual column names. The readme will help too. Adapt the code to match your data.

Q: How many families should I show in my pie chart?
A: 8-12 categories maximum. More than that becomes hard to read. Group rare families into “Other”.

Q: Should I filter out plants with missing data?
A: Up to you! Justify your choice in the critique. You could exclude NAs or show them as a separate category.

Q: Can I create an interactive visualization?
A: Yes! Use Plotly (Python/R) or Tableau. Submit the HTML file as a bonus, but include screenshots in your PDF (required).

Q: What if I want to explore something not in the options?
A: Great! As long as you meet the requirements (4 proportion viz, 1 hierarchical, critique), explore what interests you.

Q: How critical should I be in my critique?
A: Very! Honest, thoughtful critique earns more points than generic praise. Identify real weaknesses and suggest improvements.

Q: Can I work with a partner?
A: Discuss ideas and help debug code, but submit your own work. Your visualizations and critique must be entirely your own.

Q: What if I find something really interesting in the data?
A: Awesome! Mention it in your critique. Curiosity and genuine insights are valued.

Q: I’ve never heard of some of these plant families. Is that okay?
A: Totally fine! You’re learning about data visualization, not botany. The patterns are what matter.

--- title: "Lab 3: Proportions and Hierarchies" subtitle: "STAT 80B: Data Visualization - Edible Plants Edition" format: html: toc: true toc-depth: 3 code-fold: show code-tools: true --- ## Overview **Due:** Thursday, at the end of class **Points:** 10 **Submission:** PDF with visualizations + written critique In this lab, you'll practice creating and critiquing visualizations of proportions and hierarchical data using the [**TidyTuesday Edible Plants Database**](https://github.com/rfordatascience/tidytuesday/blob/main/data/2026/2026-02-03/readme.md). You'll create multiple visualizations of the same data and then analyze which ones work best and why. ## Learning Objectives By completing this lab, you will: - Create different types of proportion visualizations (pie charts, stacked bars, side-by-side bars) - Build hierarchical visualizations (treemaps, sunburst charts) - Critically evaluate the strengths and weaknesses of different visualization choices - Develop design judgment for choosing appropriate visualizations - Explore patterns in edible plant diversity across families, regions, and uses --- ## Dataset: Edible Plants Database For this lab, you'll use the **TidyTuesday Edible Plants Database** (February 3, 2026, Week 5). ### About the Data This dataset contains information about edible plants from around the world, including: - **Plant families:** Taxonomic classification (e.g., Fabaceae, Solanaceae, Rosaceae) - **Plant cultivation:** Cultivation class. Brassica, Legume, etc. - **Sun light:** How much sunlight the plant requires. - **Common and scientific names:** Plant identification - Possibly: nutritional information, culinary uses, growing conditions ### Loading the Data ::: {.panel-tabset} ## R ```r # Install tidytuesdayR if needed # install.packages("tidytuesdayR") # Load required packages library(tidyverse) library(treemapify) # For treemaps # Load data directly from TidyTuesday tuesdata <- tidytuesdayR::tt_load('2026-02-03') # OR tuesdata <- tidytuesdayR::tt_load(2026, week = 5) edible_plants <- tuesdata$edible_plants # View structure glimpse(edible_plants) head(edible_plants) names(edible_plants) # See all column names # View the readme for more information about the dataset readme(tuesdata) ``` ## Python ```python # Import required libraries import pandas as pd import matplotlib.pyplot as plt import seaborn as sns import squarify # For treemaps: pip install squarify import plotly.express as px # For interactive plots # Set style sns.set_style("whitegrid") plt.rcParams['figure.dpi'] = 300 # Load data edible_plants = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2026/2026-02-03/edible_plants.csv') # View structure print(edible_plants.head()) print(edible_plants.info()) print(edible_plants.columns) print(edible_plants.describe()) ``` ## Tableau 1. Open Tableau Desktop 2. Download CSV from GitHub `https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2026/2026-02-03/edible_plants.csv` - In Tableau: Connect to Data → Text file → select downloaded CSV **Another option - Use R to download:** - Run R code from the R section - Export: `write_csv(edible_plants, "edible_plants.csv")` - Load into Tableau ::: ### Exploring the Data **First step:** Explore what's actually in your dataset! ```r # In R - explore the structure glimpse(edible_plants) summary(edible_plants) # Check unique values for key columns edible_plants %>% count(family) %>% arrange(desc(n)) edible_plants %>% count(plant_type) %>% arrange(desc(n)) # or whatever the column is called ``` In every tool: explore all the variables, how many of them are categorical, how many are numerical, what do they represent, and which questions could you answer using them? --- ## Part 1: Visualizing Plant Diversity Proportions (4 points) ### Task Create **FOUR different visualizations** showing compositional patterns in the edible plants data. **Choose ONE of these approaches:** #### Option A: Plant Family Distribution Analyze which plant families contribute the most edible species. Create these four visualizations: 1. **Pie chart** - showing proportion of plants by family (top 8-10 families) 2. **Bar chart** - ordered by count (most common families first) 3. **Bar chart** - different ordering (alphabetical, or grouped by botanical order) 4. **Donut chart** - alternative circular visualization with top families highlighted **Example questions to explore:** - What proportion of edible plants come from the Fabaceae (legume) family? - How dominant is the Rosaceae family (roses, apples, cherries)? - Are edible plants concentrated in a few families or widely distributed? #### Option B: Plant Type Comparison Across Water requirement Compare the distribution of plant types across different categories of how much water the plant requires. Create these four visualizations: 1. **Stacked bar chart** - comparing plant types (vegetables, fruits, grains, herbs) by water requirement 2. **100% stacked bar chart** - proportional distribution showing composition differences 3. **Side-by-side bar chart** - for direct comparison of each type across categories 4. **Small multiples** - separate pie or bar charts for each categories **Example questions to explore:** - Does fruit species need more water? - Which water requirement level have the greatest diversity of edible plants? - How does plant type composition vary by water requirements? ### Requirements Each visualization must: - ✅ Be properly labeled (title, axis labels, legend if needed) - ✅ Show counts or percentages clearly - ✅ Use appropriate colors (consider green palette for plants, or color by category) - ✅ Be readable (appropriate size, font, spacing) - ✅ Include a brief caption (2-3 sentences) explaining what it shows and **one key insight** ### Data Preparation Tips You'll need to: - **Count occurrences**: Use `count()` in R or `value_counts()` in Python - **Calculate percentages**: `count / total * 100` - **Filter to top categories**: Too many small slices make pie charts unreadable - Consider showing only top 10-12 families/categories - Group remaining into "Other" - **Handle missing data**: Decide how to deal with NAs - **Create clear labels**: Shorten long scientific names if needed ### Deliverable - 4 visualizations (can be arranged using Dashboard on Tableau) - Each with a 2-3 sentence caption that includes one specific insight - Professional appearance with consistent styling --- ## Part 2: Visualizing Plant Classification Hierarchies (4 points) ### Task Create **ONE hierarchical visualization** showing the nested structure of edible plant classifications. ### Recommended Hierarchy Structures **Option 1: Taxonomic Hierarchy** - **Level 1:** Taxonomy (e.g., Fabaceae, Rosaceae, Solanaceae) - **Level 2:** Individual species count OR specific edible parts *Best for:* Understanding botanical diversity and which families dominate our food system **Option 2: Two-Level Simplified** - **Level 1:** Taxonomy (top 15-20 families) - **Level 2:** Water needs within each family *Best for:* Clear, focused visualization when you have limited space ### Visualization Type Options Choose ONE: - **Treemap** (recommended) - rectangles sized by count, excellent for comparing categories - **Sunburst chart** - circular rings showing hierarchy, good for part-whole relationships - **Icicle plot** - vertical rectangles, like a treemap rotated - **Circle packing** - nested circles, aesthetically pleasing but harder to compare sizes ### Requirements Your visualization must: - ✅ Show at least 2 levels of hierarchy clearly - ✅ Use size to encode importance (larger = more species, more diversity) - ✅ Use color strategically: - Color by plant family (with consistent palette) - OR color by plant type (vegetables=green, fruits=red/orange, etc.) - OR color by region - NOT random colors! - ✅ Label major categories clearly (family names, major groups) - ✅ Include a clear, descriptive title - ✅ Have a 2-3 sentence caption explaining the main pattern or insight ### Deliverable - 1 hierarchical visualization - Professional appearance with thoughtful design - Caption explaining what patterns are visible --- ## Part 3: Critical Evaluation (2 points) ### Task Write a **2-page critique** (approximately 200-300 words) analyzing your visualizations. Address the following questions: #### Section 1: Comparing Proportion Visualizations For your four proportion visualizations: 1. **Which visualization is most effective for showing water requirement diversity patterns?** 2. **Which visualization is least effective?** 3. **For what audience or purpose would each be appropriate?** 4. **What design choices did you make and why?** #### Section 2: Evaluating Your Hierarchical Visualization For your hierarchical visualization: 1. **What story does it tell about edible plant diversity?** 2. **What works well in your design?** 3. **What could be improved?** 4. **Would a different visualization type be better?** ### Deliverable - Written analysis - Thoughtful, specific critique with concrete examples from your visualizations - References to visualization principles from class (pre-attentive attributes, color theory, etc.) - Professional writing (well-organized, proofread!) --- ## Grading Rubric ### Part 1: Proportions (4 points) | Criterion | Points | What We're Looking For | |-----------|--------|------------------------| | Four distinct visualizations | 1.6 | All four types completed; appropriate for botanical/count data; different enough to compare | | Design quality | 1.2 | Professional appearance; thoughtful color choices; proper sizing; clean layout | | Labels and annotations | 0.8 | Clear titles, axis labels, legends; counts/percentages shown appropriately; readable text | | Captions with insights | 0.4 | 2-3 sentence captions that explain the visualization AND state one specific insight about plants | ### Part 2: Hierarchy (4 points) | Criterion | Points | What We're Looking For | |-----------|--------|------------------------| | Appropriate visualization choice | 1 | Chosen type suits the plant data and classification hierarchy | | Hierarchy clearly shown | 1 | Multiple levels visible; parent-child relationships clear; well-organized structure | | Design quality | 1 | Effective use of size (species count), color (families/types), layout; professional look | | Labels and annotations | 1 | Major categories labeled; clear title; insightful 2-3 sentence caption about diversity patterns | ### Part 3: Critique (2 points) | Criterion | Points | What We're Looking For | |-----------|--------|------------------------| | Comparison of proportion visualizations | 0.8 | Thoughtful comparison; specific examples; considers different audiences; identifies trade-offs | | Evaluation of hierarchical visualization | 0.8 | Honest self-assessment; identifies both strengths and weaknesses; considers alternatives | | Writing quality | 0.4 | Clear, well-organized; proper grammar; professional tone; references class concepts | **Total: 10 points** --- ## Technical Instructions & Code Examples ### R Code Examples #### Data Preparation & Exploration ```r library(tidyverse) library(treemapify) # Load data tuesdata <- tidytuesdayR::tt_load('2026-02-03') edible_plants <- tuesdata$edible_plants # FIRST: Explore your data! glimpse(edible_plants) head(edible_plants, 20) names(edible_plants) # Look at unique values for key columns # (Adjust column names based on what you actually have!) edible_plants %>% count(family, sort = TRUE) edible_plants %>% count(plant_type, sort = TRUE) # Option 1: Count plants by family family_counts <- edible_plants %>% count(family, sort = TRUE) %>% mutate(percentage = n / sum(n) * 100) # Filter to top 10 families for cleaner visualizations top_families <- family_counts %>% slice_max(n, n = 10) # Add "Other" category for remaining families top_families_with_other <- family_counts %>% mutate(family_group = if_else(family %in% top_families$family, family, "Other")) %>% group_by(family_group) %>% summarize(n = sum(n)) %>% mutate(percentage = n / sum(n) * 100) # Option 2: Count by edible part edible_part_counts <- edible_plants %>% count(edible_part, sort = TRUE) %>% # Adjust column name! mutate(percentage = n / sum(n) * 100) # Option 3: Cross-tabulation - families by plant type family_type <- edible_plants %>% count(family, plant_type) %>% group_by(family) %>% mutate(total = sum(n)) %>% ungroup() %>% filter(total >= 5) # Only families with 5+ species # Option 4: Regional comparison (if you have region data) regional_comparison <- edible_plants %>% count(region, plant_type) %>% group_by(region) %>% mutate(percentage = n / sum(n) * 100) ``` #### Visualization 1: Pie Chart ```r # Basic pie chart - top plant families ggplot(top_families, aes(x = "", y = n, fill = family)) + geom_bar(stat = "identity", width = 1) + coord_polar("y", start = 0) + theme_void() + labs(title = "Distribution of Edible Plants by Family (Top 10)", subtitle = "Based on TidyTuesday Edible Plants Database", fill = "Plant Family") + scale_fill_brewer(palette = "Set3") + # or "Greens", "YlGn" geom_text(aes(label = paste0(round(percentage, 1), "%")), position = position_stack(vjust = 0.5), size = 3, color = "black") ggsave("pie_chart_families.png", width = 8, height = 6, dpi = 300) # Alternative: Pie chart with "Other" category ggplot(top_families_with_other, aes(x = "", y = n, fill = family_group)) + geom_bar(stat = "identity", width = 1) + coord_polar("y", start = 0) + theme_void() + labs(title = "Edible Plant Families", fill = "Family") + scale_fill_viridis_d(option = "G", begin = 0.2, end = 0.9) ggsave("pie_chart_with_other.png", width = 8, height = 6, dpi = 300) ``` #### Visualization 2: Bar Chart (Horizontal) ```r # Horizontal bar chart - easier to read family names ggplot(top_families, aes(x = n, y = reorder(family, n), fill = family)) + geom_col(show.legend = FALSE) + labs(title = "Top 10 Edible Plant Families by Species Count", subtitle = "Number of edible species per plant family", x = "Number of Species", y = "Plant Family") + theme_minimal() + scale_fill_brewer(palette = "Set3") + geom_text(aes(label = paste0(n, " (", round(percentage, 1), "%)")), hjust = -0.1, size = 3.5) + scale_x_continuous(expand = expansion(mult = c(0, 0.15))) # Make room for labels ggsave("bar_chart_horizontal.png", width = 10, height = 6, dpi = 300) ``` #### Visualization 3: Bar Chart (Vertical, Different Order) ```r # Vertical bar chart - alphabetical order ggplot(top_families, aes(x = reorder(family, family), y = n, fill = family)) + geom_col(show.legend = FALSE) + labs(title = "Top Plant Families - Alphabetical Order", x = "Plant Family", y = "Number of Edible Species") + theme_minimal() + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 10)) + scale_fill_viridis_d(option = "viridis") + geom_text(aes(label = n), vjust = -0.5, size = 3) ggsave("bar_chart_alphabetical.png", width = 10, height = 6, dpi = 300) ``` #### Visualization 4: Stacked or Grouped Bars ```r # If comparing across regions or types: # Stacked bar chart ggplot(regional_comparison, aes(x = region, y = n, fill = plant_type)) + geom_bar(stat = "identity", position = "stack") + labs(title = "Edible Plant Types by Region", x = "Region", y = "Number of Species", fill = "Plant Type") + theme_minimal() + scale_fill_brewer(palette = "Set2") + theme(axis.text.x = element_text(angle = 45, hjust = 1)) ggsave("stacked_bar_regional.png", width = 10, height = 6, dpi = 300) # 100% stacked bar ggplot(regional_comparison, aes(x = region, y = percentage, fill = plant_type)) + geom_bar(stat = "identity", position = "fill") + labs(title = "Composition of Edible Plant Types by Region", x = "Region", y = "Percentage", fill = "Plant Type") + scale_y_continuous(labels = scales::percent_format(scale = 1)) + theme_minimal() + scale_fill_brewer(palette = "Set2") ggsave("stacked_bar_percent.png", width = 10, height = 6, dpi = 300) # Grouped (side-by-side) bars ggplot(regional_comparison, aes(x = region, y = n, fill = plant_type)) + geom_bar(stat = "identity", position = "dodge") + labs(title = "Edible Plant Types by Region - Direct Comparison", x = "Region", y = "Number of Species", fill = "Plant Type") + theme_minimal() + scale_fill_brewer(palette = "Set2") + theme(axis.text.x = element_text(angle = 45, hjust = 1)) ggsave("grouped_bar.png", width = 12, height = 6, dpi = 300) ``` #### Hierarchical Visualization: Treemap ```r # Treemap: Family > Plant Type > Count # Prepare hierarchical data treemap_data <- edible_plants %>% count(family, plant_type) %>% filter(n >= 2) # Only include family-type combos with 2+ species # Basic treemap ggplot(treemap_data, aes(area = n, fill = family, label = plant_type, subgroup = family)) + geom_treemap() + geom_treemap_text(colour = "white", place = "centre", grow = FALSE, reflow = TRUE, size = 10) + geom_treemap_subgroup_border(colour = "white", size = 3) + geom_treemap_subgroup_text(place = "centre", alpha = 0.7, colour = "white", fontface = "bold", grow = TRUE, size = 14) + labs(title = "Edible Plant Diversity: Family → Type", subtitle = "Size represents number of edible species", fill = "Plant Family") + theme(legend.position = "bottom", legend.text = element_text(size = 8)) + scale_fill_viridis_d(option = "G", begin = 0.1, end = 0.9) + guides(fill = guide_legend(nrow = 2)) ggsave("treemap_family_type.png", width = 14, height = 10, dpi = 300) # Alternative: Just families (simpler) family_counts_filtered <- family_counts %>% filter(n >= 3) # Only families with 3+ species ggplot(family_counts_filtered, aes(area = n, fill = n, label = paste0(family, "\n", n))) + geom_treemap() + geom_treemap_text(colour = "white", place = "centre", size = 12, grow = TRUE) + labs(title = "Edible Plant Families by Species Count", subtitle = "Each rectangle represents a plant family, sized by number of edible species") + scale_fill_gradient(low = "#d4f1d4", high = "#0d5c0d", name = "Species\nCount") + theme(legend.position = "right") ggsave("treemap_families_only.png", width = 12, height = 8, dpi = 300) ``` #### Creating Multi-Panel PDFs ```r # Save all plots to one PDF using gridExtra library(gridExtra) p1 <- # your pie chart ggplot object p2 <- # your first bar chart p3 <- # your second bar chart p4 <- # your stacked/grouped bar chart # Arrange in grid grid.arrange(p1, p2, p3, p4, ncol = 2, top = "Lab 3: Edible Plants Proportion Visualizations") ggsave("all_proportion_viz.pdf", width = 14, height = 10) # Alternative: patchwork package library(patchwork) (p1 | p2) / (p3 | p4) + plot_annotation(title = 'Edible Plants Visualizations', theme = theme(plot.title = element_text(size = 16, face = "bold"))) ggsave("all_proportion_viz_patchwork.pdf", width = 14, height = 10) ``` --- ### Python Code Examples #### Data Preparation ```python import pandas as pd import matplotlib.pyplot as plt import seaborn as sns import squarify # pip install squarify import plotly.express as px # Set style sns.set_style("whitegrid") plt.rcParams['figure.dpi'] = 300 # Load data edible_plants = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2026/2026-02-03/edible_plants.csv') # Explore print(edible_plants.head()) print(edible_plants.info()) print(edible_plants.columns) # Count by family family_counts = edible_plants['family'].value_counts().reset_index() family_counts.columns = ['family', 'count'] family_counts['percentage'] = (family_counts['count'] / family_counts['count'].sum()) * 100 # Top 10 families top_families = family_counts.head(10) # Count by edible part edible_part_counts = edible_plants['edible_part'].value_counts().reset_index() edible_part_counts.columns = ['edible_part', 'count'] # Cross-tabulation family_type = pd.crosstab(edible_plants['family'], edible_plants['plant_type']) ``` #### Pie Chart ```python # Pie chart plt.figure(figsize=(10, 8)) colors = sns.color_palette('Set3', len(top_families)) plt.pie(top_families['count'], labels=top_families['family'], autopct='%1.1f%%', startangle=90, colors=colors, textprops={'fontsize': 11}) plt.title('Distribution of Edible Plants by Family (Top 10)', fontsize=16, fontweight='bold', pad=20) plt.axis('equal') plt.tight_layout() plt.savefig('pie_chart.png', bbox_inches='tight', dpi=300) plt.show() ``` #### Horizontal Bar Chart ```python # Horizontal bar chart plt.figure(figsize=(10, 8)) colors_bar = sns.color_palette('Set3', len(top_families)) plt.barh(top_families['family'], top_families['count'], color=colors_bar) plt.xlabel('Number of Species', fontsize=12) plt.ylabel('Plant Family', fontsize=12) plt.title('Top 10 Edible Plant Families by Species Count', fontsize=14, fontweight='bold') # Add count labels for i, (count, pct) in enumerate(zip(top_families['count'], top_families['percentage'])): plt.text(count + max(top_families['count'])*0.02, i, f'{count} ({pct:.1f}%)', va='center', fontsize=10) plt.gca().invert_yaxis() # Highest at top plt.tight_layout() plt.savefig('bar_chart_horizontal.png', dpi=300) plt.show() ``` #### Stacked Bar Chart ```python # Stacked bar chart (example with plant types by region) # First create a pivot table pivot_data = edible_plants.groupby(['region', 'plant_type']).size().unstack(fill_value=0) # Plot ax = pivot_data.plot(kind='bar', stacked=True, figsize=(12, 6), colormap='Set2') plt.xlabel('Region', fontsize=12) plt.ylabel('Number of Species', fontsize=12) plt.title('Edible Plant Types by Region', fontsize=14, fontweight='bold') plt.legend(title='Plant Type', bbox_to_anchor=(1.05, 1), loc='upper left') plt.xticks(rotation=45, ha='right') plt.tight_layout() plt.savefig('stacked_bar.png', dpi=300) plt.show() # 100% stacked pivot_pct = pivot_data.div(pivot_data.sum(axis=1), axis=0) * 100 ax = pivot_pct.plot(kind='bar', stacked=True, figsize=(12, 6), colormap='Set2') plt.ylabel('Percentage', fontsize=12) plt.title('Composition of Plant Types by Region (%)', fontsize=14, fontweight='bold') plt.legend(title='Plant Type', bbox_to_anchor=(1.05, 1)) plt.tight_layout() plt.savefig('stacked_bar_percent.png', dpi=300) plt.show() ``` #### Treemap ```python # Static treemap using squarify plt.figure(figsize=(14, 10)) # Prepare data sizes = top_families['count'].tolist() labels = [f"{row['family']}\n{row['count']} species" for _, row in top_families.iterrows()] colors_tree = sns.color_palette('Greens', len(top_families)) squarify.plot(sizes=sizes, label=labels, color=colors_tree, alpha=0.8, text_kwargs={'fontsize': 11, 'weight': 'bold', 'color': 'white'}) plt.title('Edible Plant Families by Species Count', fontsize=16, fontweight='bold') plt.axis('off') plt.tight_layout() plt.savefig('treemap.png', dpi=300) plt.show() # Interactive treemap using Plotly # Create hierarchical data tree_data = edible_plants.groupby(['family', 'plant_type']).size().reset_index(name='count') tree_data = tree_data[tree_data['count'] >= 2] # Filter small groups fig = px.treemap(tree_data, path=['family', 'plant_type'], values='count', title='Edible Plant Diversity: Family → Type', color='family', color_discrete_sequence=px.colors.qualitative.Set3) fig.update_layout(font=dict(size=12)) fig.write_html('treemap_interactive.html') fig.show() # Save screenshot for PDF fig.write_image('treemap_interactive.png', width=1400, height=1000) ``` #### Sunburst Chart (Interactive) ```python # Sunburst chart using Plotly fig = px.sunburst(tree_data, path=['family', 'plant_type'], values='count', title='Edible Plants: Hierarchical View', color='family', color_discrete_sequence=px.colors.qualitative.Set3) fig.update_layout(font=dict(size=12)) fig.write_html('sunburst_interactive.html') fig.show() ``` --- ### Tableau Instructions #### For Proportion Visualizations **Pie Chart:** 1. Drag `Family` to **Color** (in Marks card) 2. Drag `Family` to **Angle** (or use automatic measure) 3. Change mark type to **Pie** 4. Right-click on the worksheet → **Duplicate** to count: use `COUNT([Family])` 5. Add labels: Drag `Family` to **Label**, add percentage 6. Edit colors: Click **Color** → **Edit Colors** → choose palette **Bar Chart:** 1. Drag `Family` to **Rows** 2. Drag `CNT(Record)` or count measure to **Columns** 3. Click **Sort** descending button 4. Drag `Family` to **Color** (optional) 5. Right-click axis → **Add Reference Line** if needed **Stacked Bar:** 1. Drag `Region` to **Columns** 2. Drag `Plant Type` to **Color** 3. Drag count measure to **Rows** 4. **Analysis** → **Stack Marks** → **On** 5. For percentages: Right-click axis → **Quick Table Calculation** → **Percent of Total** #### For Treemap 1. Create hierarchy: Right-click `Family` → **Hierarchy** → **Create** 2. Add `Plant Type` to the hierarchy 3. New worksheet: Change mark type to **Square** 4. Drag hierarchy to **Detail** 5. Drag count measure to **Size** 6. Drag `Family` to **Color** 7. Drag labels to **Label** 8. Format: **Color** → **Edit Colors**; **Borders** for clarity --- ## Submission Instructions ### What to Submit Submit **ONE PDF file** containing: 1. **Part 1: Four proportion visualizations** (1-2 pages) - All four visualizations clearly visible - Each with a 2-3 sentence caption that includes one specific insight 2. **Part 2: One hierarchical visualization** (1 page) - Your treemap/sunburst/etc. - With 2-3 sentence caption about diversity patterns 3. **Part 3: Written critique** (1 pages) - Well-organized with clear sections - Professional writing **Total: 3-4 pages** ### Formatting Requirements - ✅ **PDF format only** - ✅ **Filename:** `Lab3_YourLastName.pdf` - ✅ **Readable visualizations:** Not too small! Use high DPI (300+) - ✅ **Professional appearance:** Consistent fonts, clean layout - ✅ **Pages numbered** - ✅ **Your name on first page** ### How to Submit 1. Create all visualizations and save as high-quality images (PNG, 300 DPI) 2. Write your critique in Word/Google Docs 3. Combine everything into one PDF: - Word: File → Save As → PDF - Google Docs: File → Download → PDF - Or use online PDF merger 4. **Upload to Canvas before Thursday class ends** ## Tips for Success ### General Tips - **Start early!** Explore the data first before committing to specific visualizations - **Check column names** - they may differ from examples, adapt code accordingly - **Ask questions** in office hours or on the discussion board - **Be honest** in your critique - self-reflection earns points! ### Data Exploration Tips - **Look at the readme** - it has important context about the data - **Count things first** - understand distributions before visualizing - **Filter strategically** - show top 10-15 categories, combine rare ones into "Other" - **Handle NAs** - decide if you want to show missing data or exclude it ### Visualization Tips - **Less is more:** Don't try to show everything at once - **Color matters:** Use green for plants, or color by category consistently - **Label clearly:** Plant family names can be long - consider abbreviating or using horizontal bars - **Tell a story:** What's the main message each visualization should convey? - **Be consistent:** Use same color scheme across your four proportion visualizations ### Writing Tips - **Be specific:** "The Fabaceae family dominates with 47 species (23%)" is better than "One family is very common" - **Reference concepts from class:** Pre-attentive attributes, color theory, gestalt principles - **Compare thoughtfully:** Discuss trade-offs between visualization types - **Connect to real world:** What do these patterns mean for biodiversity, food security, agriculture? - **Proofread!** Typos hurt your credibility ## Frequently Asked Questions **Q: What if column names are different from the examples?** A: Expected! Use `glimpse()`, `head()`, or `info()` to see actual column names. The readme will help too. Adapt the code to match your data. **Q: How many families should I show in my pie chart?** A: 8-12 categories maximum. More than that becomes hard to read. Group rare families into "Other". **Q: Should I filter out plants with missing data?** A: Up to you! Justify your choice in the critique. You could exclude NAs or show them as a separate category. **Q: Can I create an interactive visualization?** A: Yes! Use Plotly (Python/R) or Tableau. Submit the HTML file as a bonus, but include screenshots in your PDF (required). **Q: What if I want to explore something not in the options?** A: Great! As long as you meet the requirements (4 proportion viz, 1 hierarchical, critique), explore what interests you. **Q: How critical should I be in my critique?** A: Very! Honest, thoughtful critique earns more points than generic praise. Identify real weaknesses and suggest improvements. **Q: Can I work with a partner?** A: Discuss ideas and help debug code, but submit your own work. Your visualizations and critique must be entirely your own. **Q: What if I find something really interesting in the data?** A: Awesome! Mention it in your critique. Curiosity and genuine insights are valued. **Q: I've never heard of some of these plant families. Is that okay?** A: Totally fine! You're learning about data visualization, not botany. The patterns are what matter.