How to Create Trendline in Matplotlib

In this demo article, we will have a look at the short demonstration about trendlines and also have a look at how to create a linear trendline onto a graph in Matplotlib.

Use the polyfit() Method to Create Trendline in Matplotlib

The trendline is simply a line that we draw through the data points in the plot. To estimate the trend and we use them to try to make predictions.

Let’s start with importing the required libraries in Python.

import matplotlib.pyplot as plot
import numpy as np

We have a data vector of the temperature in Celsius and a data vector of a voltage. You could easily import it into your code if you have an excel spreadsheet, a text file, or a CSV file.

In our case, we just typed seven data points manually. We are going to plot temperature as a function of voltage.

T = [20, 30, 40, 50, 60, 75, 100]
V = [1.02, 1.53, 2.05, 2.55, 3.07, 3.56, 4.05]

Source Code:

import matplotlib.pyplot as plot

# Data vectors
T = [20, 30, 40, 50, 60, 75, 100]
V = [1.02, 1.53, 2.05, 2.55, 3.07, 3.56, 4.05]

# Plot data
plot.plot(V, T, "b*")
plot.xlabel("Voltage(V)")
plot.ylabel("Temp(C)")
plot.grid()
plot.show()

Output:

Now we’ll create a trendline. Let’s compute the trend line using the polyfit() method from numpy, and this method will create a polynomial fit line.

The polyfit() method accepts the x-axis, the y-axis, or the side of the X and Y coordinates. The third parameter accepts the order, and we’ll pass it 1, 1 obviously is linear, so 2 would be quadratic.

coeff = np.polyfit(V, T, 1)  # 1=linear

Store the first element of that coefficient into the m variable, then b is the y-intercept.

m = coeff[0]
b = coeff[1]

Plot a trendline using the following syntax. We need to call the linspace() method, and this method takes some initial and ending points with hundred data points.

The Ttrend variable stores an equation, and this equation will help create a trendline.

Vtrend = np.linspace(V[0], V[-1], 100)
Ttrend = m * Vtrend + b

Source Code:

import matplotlib.pyplot as plot
import numpy as np

# Data vectors
T = [20, 30, 40, 50, 60, 75, 100]
V = [1.02, 1.53, 2.05, 2.55, 3.07, 3.56, 4.05]

# Plot data
plot.plot(V, T, "b*")
plot.xlabel("Voltage(V)")
plot.ylabel("Temp(C)")
plot.grid()

# Compute the trendline
coeff = np.polyfit(V, T, 1)  # 1=linear

m = coeff[0]
b = coeff[1]

Vtrend = np.linspace(V[0], V[-1], 100)
Ttrend = m * Vtrend + b
plot.plot(Vtrend, Ttrend, "r")
plot.show()

Output: