Simple plotting in r and python • Adam Coates

It’s time to move away from matlab entirely

While MATLAB served me well for a number of years I feel now is the time to completely stop using it all together (except for exceptional use cases, for example Psychtoolbox, but even then I’m much more inclined to use the open source MATLAB version Octave).

MATLAB themselves even write about the differences between MATLAB and Python and which one you should use. It’s true MATLAB’s documentation on functions and packages is really extensive, easily understandable and takes the lead over Python and R. But this is just because MATLAB is a private company and the subscriptions aren’t exactly cheap. For MATLAB, they argue that you pay for more than just the GUI and software but also this user support (although I’ve never ever picked up the phone and called MATLAB directly with a problem I’ve had).

Still Python has, in recent years, closed the gap between computational speed of large matrices and with an ever increasing user base, has gained a lot of very useful packages and modules.

Even R is/ (arguably has) already closed this gap too. And with Rstudios nicely laid out default GUI there is really no excuse even for beginners not to use it.

MATLAB will always be used in certain fields though because it really just works well, not to mention how easy it is to use (again though the extensive and well-documented functions are partly the reason for this).

But R and Python are open source and FREE! Anyone can use them.

I want to make plots in an open source way, how?

This blog post won’t give a full tutorial, just check out a YouTube video for that, there’s really loads! Or even ask chatgpt to do it for you, since it seems to do well at basic coding anyway.

With a fully configured neovim setup (like the one I have) it makes coding easier and much faster to do. This is an additional reason to move away from MATLAB since you almost have to do everything inside MATLABs gui.

Bar plots

Here’s an example bar plot in R using ggplot:

#| fig-align: center

library(ggplot2)

x <- runif(5, min = 1, max = 10)
fruits <- c("Apples", "Bananas",
  "Oranges", "Grapes", "Pears")

fruit_data <- data.frame(fruit = fruits, quantity = x)

ggplot(fruit_data, aes(x = fruit, y = quantity)) +
  geom_bar(stat = "identity", fill = "skyblue") +
  labs(x = "Fruit", y = "Quantity",
    title = "Amount of Fruit")

In MATLAB creating the same plot would take quite a few extra lines of code for sure.

Here’s an example bar plot in python using plt.bar:

#| fig-align: center
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np

# Generate random data
np.random.seed(0)
x = np.random.uniform(1, 10, 5)
fruits = ["Apples", "Bananas",
          "Oranges", "Grapes", "Pears"]

# Create a DataFrame
fruit_data = pd.DataFrame({"fruit": fruits, "quantity": x})

# Create the bar plot
plt.bar(fruit_data['fruit'],
        fruit_data['quantity'],
        color='skyblue')
plt.xlabel('Fruit')
plt.ylabel('Quantity')
plt.title('Amount of Fruit')
plt.show()

Scatter plots

Here are some examples of scatter plots using python:

#| fig-align: center
import matplotlib.pyplot as plt
import numpy as np

# Sample data
np.random.seed(0)
x = np.random.uniform(1, 10, 5)
fruits = ["Apples", "Bananas", "Oranges", "Grapes", "Pears"]

# Create scatter plot
plt.figure(figsize=(8, 5))
plt.scatter(fruits, x,
            color='skyblue',
            s=100,
            edgecolor='black',
            alpha=0.7)

# Add labels and title
plt.xlabel('Fruit')
plt.ylabel('Quantity')
plt.title('Amount of Fruit')

# Rotate x-axis labels for better readability
plt.xticks(rotation=45)

# Display the plot
plt.grid(True)
plt.tight_layout()
plt.show()

Here’s a scatter plot in python showing the line of best fit by just plotting y = mx + b:

#| fig-align: center
import matplotlib.pyplot as plt
import numpy as np

# Generate random data
np.random.seed(0)
x = np.random.rand(20)
y = np.random.rand(20)

# Calculate the line of best fit
m, b = np.polyfit(x, y, 1)

# Calculate correlation
correlation = np.corrcoef(x, y)[0, 1]

# Create scatter plot
plt.figure(figsize=(8, 5))
plt.scatter(x, y, color='skyblue', label='Data Points')

# Plot the line of best fit
plt.plot(x, m*x + b, color='red', label='Line of Best Fit')

# Add labels and title
plt.xlabel('X')
plt.ylabel('Y')
plt.title('Scatter Plot with Line of Best Fit')

# Add legend
plt.legend()

# Calculate and print correlation
print('Correlation coefficient:', correlation)

# Display the plot
plt.grid(True)
plt.tight_layout()
plt.show()

Here’s the exact same scatter plots but in R instead using geom_point:

#| fig-align: center
library(ggplot2)

# Sample data
set.seed(0)
x <- runif(5, min = 1, max = 10)
fruits <- c("Apples", "Bananas",
  "Oranges", "Grapes", "Pears")

# Create data frame
fruit_data <- data.frame(fruit = fruits, quantity = x)

# Create scatter plot
scatter_plot <- ggplot(fruit_data,
  aes(x = fruit, y = quantity)) +
  geom_point(color = "skyblue", size = 5,
    shape = 19, alpha = 0.7) +
  labs(x = "Fruit", y = "Quantity",
    title = "Amount of Fruit") +
  theme(axis.text.x = element_text(angle = 45,
    hjust = 1))  # Rotate x-axis labels

# Display the plot
print(scatter_plot)

Here’s a scatter plot in R using geom_smooth to fit a line of best fit:

#| fig-align: center
# Load libraries
library(ggplot2)

# Set seed for reproducibility
set.seed(0)

# Generate random data
x <- runif(20)
y <- runif(20)

# Create data frame
data <- data.frame(x = x, y = y)

# Create scatter plot
scatter_plot <- ggplot(data, aes(x = x, y = y)) +
  geom_point(color = "skyblue") +
  labs(x = "X", y = "Y",
    title = "Scatter Plot with Line of Best Fit")

# Add line of best fit
scatter_plot <- scatter_plot +
  geom_smooth(method = "lm",
    se = FALSE,
    color = "red")

# Calculate correlation coefficient
correlation <- cor(x, y)

# Print correlation coefficient
cat("Correlation coefficient:", correlation, "\n")

# Display plot
print(scatter_plot)

This is all there is to it. So the main trick here is to get the data in a format that can be read correctly. The sure fire way is to make sure the data provided to the plots is as a type array. But of course it’s possible to index for example a pandas dataframe as an array.

What about error bars?

Error bars a crucial in any plot and really tell the more about what the plot means. In any scientific work no error bars is just a no-go.

Here I show a scatter plot in R with shading using geom_smooth just by setting se = TRUE:

#| fig-align: center
# Load library
library(ggplot2)

# Set seed for reproducibility
set.seed(0)

# Generate random data
x <- runif(20)
y <- runif(20)

# Create data frame
data <- data.frame(x = x, y = y)

# Calculate standard deviation of y
std_dev_y <- sd(y)

# Create scatter plot
scatter_plot <- ggplot(data, aes(x = x, y = y)) +
  geom_point(color = "skyblue") +
  labs(x = "X", y = "Y",
    title = "Scatter Plot with Line of Best Fit")

# Add line of best fit with shading for standard deviation
scatter_plot <- scatter_plot +
  geom_smooth(method = "lm", se = TRUE,
    fill = "skyblue", color = "red") +
  annotate("text", x = max(data$x), y = max(data$y),
           label = paste("Standard Deviation of Y:",
           round(std_dev_y, 2)),
           hjust = 1, vjust = 1)

# Calculate correlation coefficient
correlation <- cor(x, y)

# Print correlation coefficient
cat("Correlation coefficient:", correlation, "\n")

# Display plot
print(scatter_plot)

And here’s a bar plot in python with error bars by setting yerr:

#| fig-align: center
import matplotlib.pyplot as plt
import numpy as np

# Sample data
fruits = ["Apples", "Bananas", "Oranges", "Grapes", "Pears"]
quantities = np.random.randint(5, 20, size=5)
errors = np.random.randint(5,10,size=5) / 2

# Create the bar plot
plt.bar(fruits, quantities, color='skyblue',
        yerr=np.abs(errors), capsize=5)

# Add labels and title
plt.xlabel('Fruit')
plt.ylabel('Quantity')
plt.title('Amount of Fruit with Error Bars')

# Rotate x-axis labels for better readability
plt.xticks(rotation=45)

# Display the plot
plt.grid(True)
plt.tight_layout()
plt.show()

Plotting is easy

I hope the examples above convinced you that plotting in R or python is very easy. Not to mention that the plots look visually better without much tinkering (unlike in MATLAB). These are just some examples and I have made these isolated on purpose such that you can really just copy any block and alter it to your needs.

Reactions

Comments

Loading comments...