CISC482 - Lecture07

Data Exploration

Dr. Jeremy Castagno

Class Business

Schedule

• Reading 3-2: Feb 8 @ 12PM, Wednesday
• Reading 4-1: Feb 10 @ 12PM, Friday
• Reading 4-2: Feb 15 @ 12PM, Wednesday
• HW3: Feb 15 @ Midnight, Wednesday

CS Faculty Candidate

• Razuan Hossain is here on Friday
• Please attend a meet and greet at 3:15 in SBSC 112
• Extra Credit!

Visualizing Data with *one** feature

Bar Chart

• A bar chart: groups on one axis, rectangles with heights that represent the number of samples.

Example Bar Chart

Code

species_count = df.groupby(['species'])['species'].count()
ax = sns.barplot(x=species_count.index.values, y=species_count.values)
ax.bar_label(ax.containers[0])
ax;

Numerical Features

• Sometimes we want visualize numerical features
• We are interested in showing users the variation of this feature
• Histograms
• Density Plots
• Box Plots

Histogram Bar Chart

Histogram Bar Chart

• Dividing the numerical feature into small regions and then count the number of values in each region
• Notice - axis have labels!
• Notice - bar widths are small enough that you can see the distributions shape

• What do you notice about this distribution?
• What is (roughly) the most likely flipper length

Code

ax = sns.histplot(data=df, x="flipper_length_mm");

Histogram Bar Horizontal Bar Chart

• Sometimes its better to have the bar chart grow horizontal

Code

species_count = df.groupby(['species'])['species'].count()
ax = sns.barplot(data=df, y="island", x="body_mass_g", errorbar=None)
ax.bar_label(ax.containers[0]);

Density Plot

• A plot that approximates the density function of the distribution for the feature.
• Density plots can be thought of as a smoothed histogram

Code

ax = sns.kdeplot(data=df, x="body_mass_g");

Density Plot with Histogram

Code

ax = sns.histplot(data=df, x="body_mass_g", kde=True);

Box Plot

• A visual representation of the summary:
  • minimum, maximum
  • first quartile, median, third quartiles
  • outliers

Code

ax = sns.boxplot(data=df, x="body_mass_g");

Boxen Plot

• Plots more quantiles
• Provides more information about the shape of the distribution, particularly in the tails.
• 50%, 25%, 12.5%, 6.25%, 3.13%

Code

ax = sns.boxenplot(data=df, x="body_mass_g");

Multiple Features

Two features

• We visualized a single feature uses one axis to display the feature value and another axis to display the value’s frequency
• However, what if we want to communicate or investigate the relationship between two variables?
  • Scatterplot, line plots, etc!

Example Scatter Plot Data

tips = sns.load_dataset("tips")
tips.head()

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4

Example Scatter Plot

• Every point is a data point
• Point out the best tips
• Point out the worst tips

Code

sns.relplot(data=tips, x="total_bill", y="tip", aspect=1.5);

Example Line Plot

Code

tips = sns.load_dataset("tips")
sns.relplot(data=tips, x="total_bill", y="tip", aspect=1.5, kind='line');

Combined Plot

• Data points and linear regression model

Looks can be deceiving

• Always show the points…

Categorical Features

Categorical Features

• Sometimes we have a categorical feature and see the difference between two different sets
  • Category: Man, Woman. Feature: Height
  • Category: Friday, Saturday, Sunday. Feature: Avg. Tip
  • Category: Penguin Species. Feature: Body Mass

Density Plot with Histogram, Species as Hue

Code

ax = sns.histplot(data=df, x="body_mass_g", kde=True, hue='species');

Bar Plot, Sex as Hue

Code

# Draw a nested barplot by species and sex
g = sns.catplot(
    data=df, kind="bar", 
    x="species", y="body_mass_g", hue="sex", 
    errorbar=None, alpha=0.6, dodge=False
)

Bar Plot, Sex as Hue

• What species and sex combination is the most prevalent?
• Which species has the smallest numerical difference between sexes?

Code

# Draw a nested barplot by species and sex
g = sns.catplot(
    data=df, kind="bar", 
    x="species", y="body_mass_g", hue="sex", 
    errorbar=None, alpha=0.6
)

Awesome Plots

Strip Plot

Code

g = sns.catplot(
    data=df, kind="strip", 
    y="species", x="bill_length_mm", hue="sex",  aspect=2, alpha=0.6
)

Swarm Plot

• A swarm plot is a scatter plot with points jittered off the lines for the categorical feature so the points do not overlap.
• A swarm plot is useful for small datasets, but with an increasing number of points, the plots get too wide.

Swarm Plot Example

Code

g = sns.swarmplot(
    data=df, 
    y="species", x="bill_length_mm", hue='sex', alpha=0.6
)

Violin

Code

sns.violinplot(data=df, x="bill_length_mm", y="species");

Violin Explained

Data Plotting Tools

Matplotlib

• Matplotlib is a comprehensive library for creating static, animated, and interactive visualizations in Python.
• The most popular graphic library (2D and 3D)
• Out of the box, plots are functional but don’t look too nice

Matplotlib Example

Seaborn

• Seaborn is a Python data visualization library based on matplotlib
• Nice and simple api that integrates very nicely with pandas
• Just pass it a data frame and call functions like
  • histplot
  • scatterplot
  • relplot
  • boxplot

Seaborn Examples