7. Seaborn Part 2

3. Seaborn plots (cont.)

3.3 Categorical Plots

Plot Type	Function	Shows	Best For (Chemistry) Examples	Shows Individual Points?
Box Plot	sns.boxplot	Median, IQR, outliers	Comparing distributions of pH, melting points across groups	No
Violin Plot	sns.violinplot	Distribution + KDE	Understanding shape of data like absorbance under conditions	No
Strip Plot	sns.stripplot	Raw data points	Showing individual measurements of reaction yield	Yes (may overlap)
Swarm Plot	sns.swarmplot	Adjusted data points	Detailed look at replicates like synthesis output	Yes (non-overlapping)
Bar Plot	sns.barplot	Mean + CI	Summarizing average reaction times or concentrations	No
Count Plot	sns.countplot	Frequencies	Counting how many times each ion type appears in a compound class	No

Box Plot (sns.boxplot())

• Shows the distribution of data based on quartiles.

• Displays the median, interquartile range (IQR), and outliers.

• Great for comparing distributions between chemical groups or experimental conditions

A box plot (also called a box-and-whisker plot) summarizes a dataset using five-number summary statistics:

1. Minimum (excluding outliers)

2. First Quartile (Q1) — 25th percentile

3. Median (Q2) — 50th percentile

4. Third Quartile (Q3) — 75th percentile

5. Maximum (excluding outliers)

It also flags outliers as dots outside the whiskers.

Key Concepts

Term	What It Means
Quartile	Values that divide your dataset into quarters.
IQR (Interquartile Range)	Q3 − Q1, the range of the middle 50% of the data.
Whiskers	Typically extend to the last data point within 1.5 × IQR from Q1/Q3.
Outliers	Points beyond the whiskers (outside 1.5 × IQR).

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd

# Sample data
data = pd.DataFrame({
    "buffer": ["A"] * 8 + ["B"] * 8 + ["C"] * 8,
    "pH": [6.8, 7.0, 6.9, 7.2, 6.8, 7.1, 6.7, 6.9,
           7.4, 7.3, 7.6, 7.2, 7.5, 7.3, 7.8, 7.2,
           8.1, 8.0, 7.9, 8.2, 8.0, 8.3, 8.5, 8.6]
})

# Make a box plot
sns.boxplot(x="buffer", y="pH", data=data)
plt.ylabel("pH Value")
plt.title("Distribution of pH Values Across Buffers")
plt.show()

Anatomy of a Box Plot

Element	Symbol / Abbreviation	Description
Minimum	—	Smallest value within 1.5×IQR below Q1 (start of lower whisker)
Lower Whisker	—	Extends from Q1 to the smallest non-outlier value
First Quartile	Q1	25th percentile; lower edge of the box
Median	Q2	50th percentile; line inside the box
Third Quartile	Q3	75th percentile; upper edge of the box
Upper Whisker	—	Extends from Q3 to the largest non-outlier value
Maximum	—	Largest value within 1.5×IQR above Q3 (end of upper whisker)
Interquartile Range	IQR	Q3 − Q1; range of the middle 50% of the data
Outliers	—	Data points beyond 1.5×IQR from Q1 or Q3; plotted as individual dots

Interpreting Whisker Asymmetry

Whisker Pattern	Interpretation
Long upper whisker	Right-skewed distribution (positive skew) – long tail toward larger values
Long lower whisker	Left-skewed distribution (negative skew) – long tail toward smaller values
Symmetrical whiskers	Data is fairly symmetric around the median
Skewness describes asymmetry in the data. Box plots don’t calculate skew, but long whiskers often hint at it.

import numpy as np

# Create skewed data for Buffer C
np.random.seed(0)
buffer_c_skewed = np.random.exponential(scale=0.3, size=20) + 7.5  # Right-skewed

# Construct new DataFrame
data_skewed = pd.DataFrame({
    "buffer": ["A"] * 8 + ["B"] * 8 + ["C"] * 20,
    "pH": [6.8, 7.0, 6.9, 7.2, 6.8, 7.1, 6.7, 6.9,
           7.4, 7.3, 7.6, 7.2, 7.5, 7.3, 7.8, 7.2] + list(buffer_c_skewed)
})

sns.boxplot(x="buffer", y="pH", data=data_skewed)
plt.ylabel("pH Value")
plt.title("Box Plot With Skewed Buffer C")
plt.show()

Violin Plot

• Combines a box plot with a kernel density estimate (KDE).

• Visualizes the distribution shape more clearly than a box plot alone.

• Useful for highlighting bimodal or skewed data in repeated measurements

Anatomy of a Violin Plot

Feature	Description
Width of the “violin”	Represents the density of the data at that value (fatter = more data points there).
Center line	The median (like in box plot).
Inner box / bars	Can show quartiles (Q1, Q2, Q3) if enabled.
Tails	Represent the range of the data, often matching the min/max or whiskers.
Symmetry	Typically mirrored left and right, but can be customized (e.g., split violins).

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np

# Skewed buffer C for demonstration
np.random.seed(0)
buffer_c_skewed = np.random.exponential(scale=0.3, size=20) + 7.5

data_skewed = pd.DataFrame({
    "buffer": ["A"] * 8 + ["B"] * 8 + ["C"] * 20,
    "pH": [6.8, 7.0, 6.9, 7.2, 6.8, 7.1, 6.7, 6.9,
           7.4, 7.3, 7.6, 7.2, 7.5, 7.3, 7.8, 7.2] + list(buffer_c_skewed)
})

# Violin plot
sns.violinplot(x="buffer", y="pH", data=data_skewed, inner="quartile")  # inner="quartile" shows box stats
plt.title("Violin Plot of pH by Buffer (with Skew)")
plt.ylabel("pH Value")
plt.show()

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np

# Generate skewed Buffer C
np.random.seed(0)
buffer_c_skewed = np.random.exponential(scale=0.3, size=20) + 7.5

# Combine with normal data
data_skewed = pd.DataFrame({
    "buffer": ["A"] * 8 + ["B"] * 8 + ["C"] * 20,
    "pH": [6.8, 7.0, 6.9, 7.2, 6.8, 7.1, 6.7, 6.9,
           7.4, 7.3, 7.6, 7.2, 7.5, 7.3, 7.8, 7.2] + list(buffer_c_skewed)
})

# Plot side-by-side
fig, axes = plt.subplots(1, 2, figsize=(12, 5), sharey=True)

# Box plot
sns.boxplot(x="buffer", y="pH", data=data_skewed, ax=axes[0])
axes[0].set_title("Box Plot of pH by Buffer")
axes[0].set_ylabel("pH Value")

# Violin plot
sns.violinplot(x="buffer", y="pH", data=data_skewed, inner="quartile", ax=axes[1])
axes[1].set_title("Violin Plot of pH by Buffer")
axes[1].set_ylabel("")

plt.tight_layout()
plt.show()

Comparing Box to Violin Plots

Feature	Box Plot	Violin Plot
Median	Line in center of box	Same line, shown if inner='quartile'
Quartiles	Box edges	Shown inside if inner='quartile'
Whiskers	Lines to min/max within 1.5×IQR	Not explicitly shown
Outliers	Dots beyond whiskers	Not shown unless overlaid
Distribution Shape	Not shown	Clearly shown via width of violin
Skewness	Implied from whiskers/outliers	Directly visible from asymmetric violin shape

Strip Plot

• Plots individual data points along a categorical axis.

• Can show raw data for transparency but may overlap.

• Often used in conjunction with box plots for validation. Jittering means adding a small random displacement along the categorical axis (i.e., sideways) to prevent overlapping dots.

This is important because if multiple data points have the same or similar values, they would overlap on a strip plot. With jitter:

• You can see repeated values

• The plot becomes more readable

sns.stripplot(x="buffer", y="pH", data=data_skewed, jitter=True)

You can control how much jitter with jitter=0.1, 0.2, etc.

import seaborn as sns
import matplotlib.pyplot as plt

# Strip plot with jitter
sns.stripplot(x="buffer", y="pH", data=data_skewed, jitter=0.2, size=6)
plt.title("Strip Plot of pH by Buffer (with Jitter)")
plt.ylabel("pH Value")
plt.show()

Swarm Plot

• Like a strip plot, but adjusts point positions to avoid overlaps.

• Useful for dense but non-overlapping point visualization.

• Perfect for small sample sizes, e.g., yields from multiple chemical syntheses.

Feature	Strip Plot	Swarm Plot
Overlap	Can have overlapping points	Avoids overlapping points
Readability	Can become cluttered for dense data	More readable when lots of points exist
Display style	Jittered vertically or horizontally	Uses a special algorithm to arrange points
Performance	Faster with large datasets	Slower with large datasets

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
import os

# Load PubChem periodic table data
base_data_dir = os.path.expanduser("~/data")
pubchem_data_dir = os.path.join(base_data_dir, "pubchem_data")
periodictable_csv_datapath = os.path.join(pubchem_data_dir, "PubChemElements_all.csv")
df = pd.read_csv(periodictable_csv_datapath, index_col=1)
print(df.head())
# Drop rows with missing MeltingPoint or GroupBlock
df = df.dropna(subset=["MeltingPoint", "GroupBlock"])

# Create side-by-side plots
fig, axes = plt.subplots(1, 2, figsize=(14, 6), sharey=True)

# Strip Plot
sns.stripplot(data=df, x="GroupBlock", y="MeltingPoint", jitter=True, ax=axes[0])
axes[0].set_title("Strip Plot: Melting Point by Group Block")
axes[0].set_xlabel("Group Block")
axes[0].set_ylabel("Melting Point (K)")

# Swarm Plot
sns.swarmplot(data=df, x="GroupBlock", y="MeltingPoint", ax=axes[1])
axes[1].set_title("Swarm Plot: Melting Point by Group Block")
axes[1].set_xlabel("Group Block")
axes[1].set_ylabel("")

# Rotate x-axis labels for both plots
for ax in axes:
    ax.tick_params(axis='x', rotation=45)

plt.tight_layout()
plt.show()

        AtomicNumber       Name  AtomicMass CPKHexColor ElectronConfiguration  \
Symbol                                                                          
H                  1   Hydrogen    1.008000      FFFFFF                   1s1   
He                 2     Helium    4.002600      D9FFFF                   1s2   
Li                 3    Lithium    7.000000      CC80FF               [He]2s1   
Be                 4  Beryllium    9.012183      C2FF00               [He]2s2   
B                  5      Boron   10.810000      FFB5B5           [He]2s2 2p1   

        Electronegativity  AtomicRadius  IonizationEnergy  ElectronAffinity  \
Symbol                                                                        
H                    2.20         120.0            13.598             0.754   
He                    NaN         140.0            24.587               NaN   
Li                   0.98         182.0             5.392             0.618   
Be                   1.57         153.0             9.323               NaN   
B                    2.04         192.0             8.298             0.277   

       OxidationStates StandardState  MeltingPoint  BoilingPoint   Density  \
Symbol                                                                       
H               +1, -1           Gas         13.81         20.28  0.000090   
He                   0           Gas          0.95          4.22  0.000179   
Li                  +1         Solid        453.65       1615.00  0.534000   
Be                  +2         Solid       1560.00       2744.00  1.850000   
B                   +3         Solid       2348.00       4273.00  2.370000   

                  GroupBlock YearDiscovered  
Symbol                                       
H                   Nonmetal           1766  
He                 Noble gas           1868  
Li              Alkali metal           1817  
Be      Alkaline earth metal           1798  
B                  Metalloid           1808  

Bar Plot

• Plots means (or other summary statistics) with confidence intervals.

• Commonly used to summarize group averages (e.g., average reaction rate by catalyst)

Bar Plots vs. Histograms

• | Plot Type | Good For… | Chemistry Example | |————|—————————————|——————————————–| | Bar Plot | Comparing values between categories | Ionization energy by element or group | | Histogram | Exploring data distribution/spread | How melting points are distributed |

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
import os

# Load data
base_data_dir = os.path.expanduser("~/data")
pubchem_data_dir = os.path.join(base_data_dir, "pubchem_data")
periodictable_csv_datapath = os.path.join(pubchem_data_dir, "PubChemElements_all.csv")
df = pd.read_csv(periodictable_csv_datapath, index_col=1)

# Drop missing values
df = df.dropna(subset=["IonizationEnergy"])

# Subset: first 10 elements
df_subset = df.head(10).copy()
df_subset["Symbol"] = df_subset.index  # Make symbol a column for hue

# Bar plot with palette + hue
plt.figure(figsize=(10, 6))
sns.barplot(data=df_subset, x="Symbol", y="IonizationEnergy", hue="Symbol", palette="viridis", legend=False)

plt.title("Ionization Energy of First 10 Elements")
plt.xlabel("Element Symbol")
plt.ylabel("Ionization Energy (eV)")
plt.xticks(rotation=45)

plt.tight_layout()
plt.show()

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

# Load and clean data
df = pd.read_csv(periodictable_csv_datapath, index_col=1)
df = df.dropna(subset=["MeltingPoint", "BoilingPoint"])
df_subset = df.head(5).copy()
df_subset["Symbol"] = df_subset.index

# Reshape from wide to long format
df_melted = df_subset[["Symbol", "MeltingPoint", "BoilingPoint"]].melt(id_vars="Symbol",
                                var_name="Property", value_name="Temperature (K)")

# Grouped bar plot
plt.figure(figsize=(10, 6))
sns.barplot(data=df_melted, x="Symbol", y="Temperature (K)", hue="Property")

plt.title("Melting vs Boiling Point for First 5 Elements")
plt.xlabel("Element Symbol")
plt.ylabel("Temperature (K)")
plt.xticks(rotation=45)
plt.legend(title="Property")

plt.tight_layout()
plt.show()

# Sorting the bars
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

# Load and clean data
df = pd.read_csv(periodictable_csv_datapath, index_col=1)
df = df.dropna(subset=["IonizationEnergy"])
df_subset = df.copy()
df_subset["Symbol"] = df_subset.index

# Sort by IonizationEnergy
df_sorted = df_subset.sort_values("IonizationEnergy", ascending=False).head(10)

# Bar plot
plt.figure(figsize=(10, 6))
sns.barplot(data=df_sorted, x="Symbol", y="IonizationEnergy", hue="Symbol", palette="mako")

plt.title("Top 10 Elements by Ionization Energy")
plt.xlabel("Element Symbol")
plt.ylabel("Ionization Energy (eV)")
plt.xticks(rotation=45)

plt.tight_layout()
plt.show()

Count Plot

• Shows the frequency of observations in each category.

• Ideal for counting chemical types, classifications, or categorical labels.

A count plot is just a bar plot of value counts — it counts how many times each category appears in your data.

It’s like doing df['GroupBlock'].value_counts() and then turning that into a bar chart.

sns.countplot(data=df, x="GroupBlock")
plt.xticks(rotation=45);

Categorical Multiplot (faceted)

A faceted categorical plot is a grid of subplots where each subplot shows a slice of the data based on a categorical variable (e.g., phase of matter, group block, decade discovered, etc.).

Overview of Faceted Plots

Function	Best for…	Plot Types (via kind)	Use Case Example
sns.catplot()	Categorical data	"strip", "swarm", "box", "violin", "bar", "count"	Compare melting points across phases by group block
sns.relplot()	Relational (numeric ↔ numeric)	"scatter", "line"	Plot atomic radius vs. ionization energy, faceted by state
sns.displot() **Distributions	"hist", "kde", "ecdf"	Show histogram of boiling points by phase

Common to all three:

• All use col= or row= to create facet grids.

• All create Figure-level plots (not axes-level), so they manage layout for you.

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Load data
df = pd.read_csv(periodictable_csv_datapath, index_col=1)

# Clean data: drop missing values
df = df.dropna(subset=["MeltingPoint", "GroupBlock", "StandardState"])

# Plot: box plots of melting points by group, split by standard state
g = sns.catplot(
    data=df,
    x="GroupBlock", y="MeltingPoint",
    kind="box",
    col="StandardState",
    col_wrap=2,  # Wrap into 2 columns if more than 2 categories
    height=5,
    aspect=1.3
)

g.set_xticklabels(rotation=45)
g.fig.subplots_adjust(top=0.85)
g.fig.suptitle("Melting Points by Group Block, Faceted by Standard State")
plt.show()

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Load and clean data
df = pd.read_csv(periodictable_csv_datapath, index_col=1)
df = df.dropna(subset=["MeltingPoint", "GroupBlock", "StandardState"])

# Set consistent order for x categories
group_order = df["GroupBlock"].dropna().unique().tolist()

# Create the faceted box plot
g = sns.catplot(
    data=df,
    x="GroupBlock", y="MeltingPoint",
    kind="box",
    col="StandardState",
    col_wrap=2,
    height=5,
    aspect=1.3,
    sharey=False,
    order=group_order
)

# Draw the plot to populate tick labels
g.fig.canvas.draw()

# Get last axis (bottom-right plot) — only one we want to show labels
last_ax = g.axes.flat[-1]

# Rotate and align x-tick labels on bottom plot only
for label in last_ax.get_xticklabels():
    label.set_rotation(45)
    label.set_ha('right')

last_ax.set_xlabel("Group Block")

# Remove tick labels from all other axes
for ax in g.axes.flat[:-1]:
    ax.set_xticklabels([])
    ax.set_xlabel("")

# Add a main title
g.fig.subplots_adjust(top=0.85)
g.fig.suptitle("Melting Points by Group Block, Faceted by Standard State")

plt.show()

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Load and clean data
df = pd.read_csv(periodictable_csv_datapath, index_col=1)
df = df.dropna(subset=["MeltingPoint", "GroupBlock", "StandardState"])

# Set consistent order of categories
group_order = df["GroupBlock"].dropna().unique().tolist()

# Create faceted box plots
g = sns.catplot(
    data=df,
    x="GroupBlock", y="MeltingPoint",
    kind="box",
    col="StandardState",
    col_wrap=2,
    height=5,
    aspect=1.3,
    sharey=False,
    order=group_order
)

# Rotate labels on all plots
g.fig.canvas.draw()  # Make sure tick labels exist
for ax in g.axes.flat:
    for label in ax.get_xticklabels():
        label.set_rotation(45)
        label.set_ha('right')
    ax.set_xlabel("Group Block")

# Title
g.fig.subplots_adjust(top=0.85)
g.fig.suptitle("Melting Points by Group Block, Faceted by Standard State")

plt.show()

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

# Load and clean data
df = pd.read_csv(periodictable_csv_datapath, index_col=1)
df = df.dropna(subset=["MeltingPoint", "GroupBlock", "StandardState"])

# Use consistent group order
group_order = df["GroupBlock"].dropna().unique().tolist()

# Create faceted box plots
g = sns.catplot(
    data=df,
    x="GroupBlock", y="MeltingPoint",
    kind="box",
    col="StandardState",
    col_wrap=2,
    height=5,
    aspect=1.3,
    sharey=False,
    order=group_order
)

# Rotate all x-tick labels
g.fig.canvas.draw()
for ax in g.axes.flat:
    for label in ax.get_xticklabels():
        label.set_rotation(45)
        label.set_ha('right')
    ax.set_xlabel("Group Block")

# ---------- Annotate Outliers in the 'Solid' Plot ----------

# Get the axis for the 'Solid' subplot
solid_ax = None
for ax, title in zip(g.axes.flat, g.col_names):
    if title == 'Solid':
        solid_ax = ax
        break

if solid_ax:
    # Filter solid-state data
    solid_df = df[df["StandardState"] == "Solid"]

    # Annotate outliers for each group
    for group in solid_df["GroupBlock"].unique():
        group_data = solid_df[solid_df["GroupBlock"] == group]
        y = group_data["MeltingPoint"]
        q1 = y.quantile(0.25)
        q3 = y.quantile(0.75)
        iqr = q3 - q1
        upper = q3 + 1.5 * iqr

        outliers = group_data[group_data["MeltingPoint"] > upper]

        # Get the x-position from the group order
        x_pos = group_order.index(group)
        for symbol, row in outliers.iterrows():
            solid_ax.annotate(
                text=symbol,
                xy=(x_pos, row["MeltingPoint"]),
                xytext=(5, 5),
                textcoords="offset points",
                fontsize=8,
                rotation=30,
                ha='left'
            )

# Title
g.fig.subplots_adjust(top=0.85)
g.fig.suptitle("Melting Points by Group Block, Faceted by Standard State\n(Outliers Annotated in Solid Plot)")

plt.show()

Acknowledgements

This content was developed with assistance from Perplexity AI and Chat GPT. Multiple queries were made during the Fall 2024 and the Spring 2025.