This guide will show you how to use Jupyter Notebook with the MCP2221(A) to connect I2C sensors from your desktop PC running Windows, macOS or Linux. You can use any CircuitPython library for any of our I2C sensors to stream data into your computer's USB port.

We've written three interactive Jupyter Notebooks for three different types of sensors - a temperature sensor, an accelerometer and a thermal camera. All of these notebooks have animated graphs so you can see data streaming into your computer in real-time.

This guide is also compatible with the Adafruit FT232H breakout (EXCEPT for the MLX thermal camera example). You'll need to make a small adjustment to the code. See the Jupyter Notebook Examples page for more information.

Parts

Materials

The MCP2221A has a USB-C connector, make sure you pick up the correct cable or adapter for your computer.  

The sensors we selected for this guide can be used with a STEMMA QT cable so you can plug-and-play with the MCP2221's STEMMA QT port. 

Set up MCP2221

This guide assumes you've set up your computer to interface with the MCP2221. If you have not yet set up the MCP2221 for your computer, click the link below and come back to this page once you have everything set up.

• MCP2221 Setup Page

Install Anaconda

If you're new to all this, the Jupyter Project recommends installing Anaconda. This package installs the latest stable version of Python, Jupyter Notebook, and other commonly used packages for scientific computing and data science.

Navigate to the Anaconda downloads page, select your operating system, and download the installer including Python 3.7+.

Install the version of Anaconda you downloaded by following the executable's instructions.

Launching Jupyter Notebook

Once Anaconda is installed, open the Anaconda Navigator Application.

Under Jupyter Notebook, click Launch

Jupyter Notebook should open in a new web browser at the URL http://localhost:8888/notebooks. To ensure you set up the MCP2221 correctly, we created a Jupyter Notebook.

Download all the notebooks required for this project by clicking Download: Project ZIP on the upper left hand side of the embedded preview below. 

Once downloaded, unzip the file and keep it somewhere safe like on your desktop.

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Python Software Package Installation\n",
    "import sys\n",
    "!{sys.executable} -m pip install adafruit-blinka adafruit-circuitpython-msa301 hidapi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set an Environment Variable so Adafruit Blinka knows we're using the MCP2221\n",
    "import os\n",
    "os.environ[\"BLINKA_MCP2221\"] = \"1\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Attempt to import a CircuitPython Module\n",
    "import board"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Python Software Package Installation\n",
    "import sys\n",
    "!{sys.executable} -m pip install adafruit-blinka adafruit-circuitpython-msa301 hidapi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set an Environment Variable so Adafruit Blinka knows we're using the MCP2221\n",
    "import os\n",
    "os.environ[\"BLINKA_MCP2221\"] = \"1\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Attempt to import a CircuitPython Module\n",
    "import board"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

Code Usage

Jupyter Notebooks are split into cells. Cells may contain code, images, equations or text. This example notebook only contains code cells.

The first code cell contains two lines. The first imports the sys module which imports functions that interact with the Python interpreter. The second line installs all the dependencies we need to use this notebook with the MCP2221, including adafruit-blinka and hardware support packages. 

You should get no errors at all, in which case you can continue onto the examples!

Error: Board not supported None

If you get NotImplementedError: Board not supported None,

That could mean you did not set the MCP2221 environmental variable or you don't have the latest Python libraries installed or the MCP2221 is not plugged in to USB.

Error: BLINKA_MCP2221 environment variable set, but no MCP2221 device found

If you get this error, check your USB cable - it could be that you have a charge-only not charge+sync cable. Your board may also be unplugged from USB.

Now that you have Jupyter installed and Blinka set up on your computer, let's play around!

The following pages contain code examples as downloadable Jupyter notebooks. The examples use some popular sensors we have and should serve as a jumping off point for your experimentation.

If you're looking for more sensors - check out our growing range of plug-and-play STEMMA sensors here. Most of these have CircuitPython libraries available and are compatible with this guide.

Compatibility with FT232H

The notebooks in this guide are also compatible with the Adafruit FT232H breakout and CircuitPython Libraries.

If you're using a FT232H breakout, make sure you change the BLINKA_MCP2221 environment variable to BLINKA_FT232H.

• Read the learning system guide here to get set up with the FT232H and CircuitPython Libraries...

This example is a Jupyter notebook which graphs the current temperature using a PCT2075 temperature sensor. The graph is updated in real-time with temperature values and we've added horizontal lines across the X-axis to show temperature maximum and minimum thresholds.

While this notebook is designed for the PCT2075 sensor, you can easily implement one of the many temperature sensors available on the Adafruit website. Be sure to check if it has a CircuitPython library first!

Wiring

We'll be using the PCT2075 sensor to precisely measure temperature. The MCP2221 and PCT2075 both have STEMMA QT connectors, so you can either wire it up on a breadboard or use a STEMMA QT cable.

Then, download the example notebook:

In the Jupyter file browser, click Upload. From the file browser, select the PCT2075.ipynb example.

The next cell imports CircuitPython modules (such as board and busio) and initializes the I2C connection with the sensor. To verify that your board is properly initialized, it will also print a temperature reading from the PCT2075.

Click the cell containing the code to graph the temperature sensor. The graph should update every 5 seconds with a new reading. 

Code Walkthrough

Let's walk through this notebook, cell-by-cell, to understand how this code works.

First, we import all the required libraries for the notebook. We'll be using matplotlib to plot the temperature data from our sensor. We'll also invoke the special %matplotlib notebook magic to tell matplotlib we're using a Jupyter notebook.

%matplotlib notebook
from datetime import datetime
import matplotlib.pyplot as plt
from collections import deque
from matplotlib.animation import FuncAnimation

%matplotlib notebook
from datetime import datetime
import matplotlib.pyplot as plt
from collections import deque
from matplotlib.animation import FuncAnimation

Next we declare some constants like HISTORY_SIZE (how many sensor samples we're displaying on the graph), and INTERVAL (the graph's update interval, in seconds). We'll also declare MAX_TEMP and MIN_TEMP which are used to generate horizontal lines across the x-axis for displaying the maximum and minimum temperature values.

# How many sensor samples we want to store
HISTORY_SIZE = 100

# Graph update interval (in seconds)
INTERVAL = 5

# Maximum Temperature (in degrees C)
MAX_TEMP = 30

# Minimum Temperature (in degrees C)
MIN_TEMP = 10

# How many sensor samples we want to store
HISTORY_SIZE = 100

# Graph update interval (in seconds)
INTERVAL = 5

# Maximum Temperature (in degrees C)
MAX_TEMP = 30

# Minimum Temperature (in degrees C)
MIN_TEMP = 10

Our code plots and displays 100 sensor readings at a time. We store these readings in a list-like object called a deque container datatype. If you've never seen this datatype before, don't worry - it's very similar to a list object except it's "optimized for fast fixed-length operations" and support a maxlen argument which sets the maximum possible size of a deque. When the deque grows beyond its maxlen size, it pops objects off of its opposite end (like a FIFO stack).

We use one deque to store sensor readings (temp_data) and another deque to store time-stamps (x_time)

# Global x-axis array
x_time = deque(maxlen=HISTORY_SIZE)

# Temperature data
temp_data = deque(maxlen=HISTORY_SIZE)

# Global x-axis array
x_time = deque(maxlen=HISTORY_SIZE)

# Temperature data
temp_data = deque(maxlen=HISTORY_SIZE)

Next up, let's make a new plot and give it a title. 

# Create new plot
fig, ax = plt.subplots()

# Global title
fig.suptitle("PCT2075 Temperature", fontsize=14)

# Create new plot
fig, ax = plt.subplots()

# Global title
fig.suptitle("PCT2075 Temperature", fontsize=14)

In the animate method, we'll poll the temperature sensor and store it in the temp_data deque. Using the CPython datetime module, the code takes the current time and formats it using strftime for display on the x-axis as ticks.

# Read the temperature sensor and add the value to the temp_data array
temp_data.append(pct.temperature)

# Grab the datetime, auto-range based on length of accel_x array
x_time.append(datetime.now().strftime('%M:%S'))

# Read the temperature sensor and add the value to the temp_data array
temp_data.append(pct.temperature)

# Grab the datetime, auto-range based on length of accel_x array
x_time.append(datetime.now().strftime('%M:%S'))

The next chunk of code clears the axis, constrains the y-axis to display a maximum value of 50 degrees celsius and 0 degrees celsius. It also adds a descriptive label to the Y-Axis.

# Clear axis prior to plotting
ax.cla()

# Constrain the Y-axis
plt.ylim(top=50,bottom=0)

# Y-Axis label
plt.ylabel('Temperature\n(c)')

# Clear axis prior to plotting
ax.cla()

# Constrain the Y-axis
plt.ylim(top=50,bottom=0)

# Y-Axis label
plt.ylabel('Temperature\n(c)')

We'll use the autofmt_date method to rotate and align the x-axis tick labels. Then, let's add a grid so we can see our data better.

fig.autofmt_xdate()
ax.grid(True, linestyle=':', linewidth=0.5)

fig.autofmt_xdate()
ax.grid(True, linestyle=':', linewidth=0.5)

Next up, plot the temperature graph and two dotted horizontal lines across the x-axis to represent  maximum and minimum temperature values.

# Add a horizontal minimum line across the X-axis
plt.axhline(y=MAX_TEMP, color='r', linestyle=':', label='Max. Temperature')

# Add a horizontal maximum line across the X-axis
plt.axhline(y=MIN_TEMP, color='b', linestyle=':', label='Min. Temperature')

# Add a horizontal minimum line across the X-axis
plt.axhline(y=MAX_TEMP, color='r', linestyle=':', label='Max. Temperature')

# Add a horizontal maximum line across the X-axis
plt.axhline(y=MIN_TEMP, color='b', linestyle=':', label='Min. Temperature')

Let's add a legend to the graph. This will make it easy to discern which line is which if we look back at it later, or if the notebook is shared with a colleague or friend. 

Matplotlib's ax.legend() method automatically creates a legend for your graph, provided each plot has a label attached to it.

# Add a legend to the graph
ax.legend()

# Add a legend to the graph
ax.legend()

We'll pause the plot's output for INTERVAL seconds

# Pause the plot for INTERVAL seconds 
plt.pause(INTERVAL)

# Pause the plot for INTERVAL seconds 
plt.pause(INTERVAL)

Finally, this method "makes an animation by repeatedly calling a function", animate. We provide it the fig we generated earlier and the function we'd like to animate.

ani = animation.FuncAnimation(fig, animate)

ani = animation.FuncAnimation(fig, animate)

Wiring

We'll be using the LSM6DSOX sensor to precisely measure acceleration data. The MCP2221 and LSM6DSOX both have STEMMA QT connectors, so you can either wire it up on a breadboard or use a STEMMA QT cable.

Then, download the example notebook:

Code Usage

In the Jupyter file browser, click Upload. From the file browser, select the LSM6DSOX_Accel.ipynb example.

The next cell sets an environment variable so Blinka knows we're using the MCP2221.

Then, it imports CircuitPython libraries and initializes the I2C connection with the sensor.

The next cell imports CircuitPython modules (such as board and busio) and initializes the i2c connection with the sensor. To verify that your board is properly initialized, it should also values form the LSM6DSOX's acceleration sensor.

The third code cell uses Matplotlib to generate a graph for the LSM6DSOX's acceleration data. We used three side-by-side subplots to visualize the X, Y, and Z axis. 

Increasing the Number of Sensor Readings

By default, this code cell only displays 20 sensor readings. If you want to display more sensor readings on your graph, simply change the value of the HISTORY_SIZE variable in the code cell and re-run it. 

For more information about how this code cell works, read on!

Code Walkthrough

First, we import all required libraries for this code cell. Most of the libraries come from the Matplotlib library. We use this library to plot the data obtained by our sensor.

%matplotlib notebook
import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation
import datetime
import matplotlib.dates as mdates
from collections import deque

%matplotlib notebook
import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation
import datetime
import matplotlib.dates as mdates
from collections import deque

Our code only plots and displays 20 sensor readings at a time. We store these readings in a list-like object called a deque container datatype. If you've never seen this datatype before, don't worry - it's very similar to a list object except it's "optimized for fast fixed-length operations" and support a maxlen argument which sets the maximum possible size of a deque. When the deque grows beyond its maxlen size, it pops objects off of its opposite end (like a FIFO stack).

We'll be using four deque objects to represent the x-axis, the first graph's y-axis (accelerometer's x-axis data), the second graph's y-axis (accelerometer's y-axis data), and the third graph's y-axis (accelerometer's y-axis data). These deque objects use HISTORY_SIZE as the deque's maxlen. You may increase the amount of data to display on the graph by increasing HISTORY_SIZE.

# Deque for X-Axis (time)
x_vals = deque(maxlen=HISTORY_SIZE)

# Deque for Y-Axis (accelerometer readings)
accel_x = deque(maxlen=HISTORY_SIZE)
accel_y = deque(maxlen=HISTORY_SIZE)
accel_z = deque(maxlen=HISTORY_SIZE)

# Deque for X-Axis (time)
x_vals = deque(maxlen=HISTORY_SIZE)

# Deque for Y-Axis (accelerometer readings)
accel_x = deque(maxlen=HISTORY_SIZE)
accel_y = deque(maxlen=HISTORY_SIZE)
accel_z = deque(maxlen=HISTORY_SIZE)

Next, we'll create three side-by-side sub-plots and call tight_layout to adjust the subplot parameters to give nicer padding between examples.

• For more information about spacing matplotlib subplots, check out this guide...

# Create 3 side-by-side subplots
fig, (ax1, ax2, ax3) = plt.subplots(1,3)

# Automatically adjust subplot parameters for nicer padding between plots
plt.tight_layout()

# Create 3 side-by-side subplots
fig, (ax1, ax2, ax3) = plt.subplots(1,3)

# Automatically adjust subplot parameters for nicer padding between plots
plt.tight_layout()

Let's now take a look at the animate method. This method polls the LSM303's acceleration values and stores them in a tuple named accel_data. Next, the code appends the values from the tuple to deque objects, accel_x, accel_y, accel_z.

def animate(i):
    # Poll the LSM303AGR
    accel_data = accel.acceleration
    # Add the X/Y/Z values to the accel arrays
    accel_x.append(accel_data[0])
    accel_y.append(accel_data[1])
    accel_z.append(accel_data[2])

def animate(i):
    # Poll the LSM303AGR
    accel_data = accel.acceleration
    # Add the X/Y/Z values to the accel arrays
    accel_x.append(accel_data[0])
    accel_y.append(accel_data[1])
    accel_z.append(accel_data[2])

We grab the current time (in seconds using CPython's datetime module) and store it in a deque, x_vals.

# Grab the datetime, auto-range based on length of accel_x array
x_vals = [datetime.datetime.now() + datetime.timedelta(seconds=i) for i in range(len(accel_x))]

# Grab the datetime, auto-range based on length of accel_x array
x_vals = [datetime.datetime.now() + datetime.timedelta(seconds=i) for i in range(len(accel_x))]

Now we're up to the fun part of this code walkthrough - displaying the graphs. Since we're "animating" the graph, the axis will need to be cleared and re-drawn each time the animate method runs. Let's clear the three axis, set up grid titles and enable grid lines. 

# Clear all axis
ax1.cla()
ax2.cla()
ax3.cla()

# Set grid titles
ax1.set_title('X', fontsize=10)
ax2.set_title('Y', fontsize=10)
ax3.set_title('Z', fontsize=10)

# Enable subplot grid lines
ax1.grid(True, linewidth=0.5, linestyle=':')
ax2.grid(True, linewidth=0.5, linestyle=':')   
ax3.grid(True, linewidth=0.5, linestyle=':')

# Clear all axis
ax1.cla()
ax2.cla()
ax3.cla()

# Set grid titles
ax1.set_title('X', fontsize=10)
ax2.set_title('Y', fontsize=10)
ax3.set_title('Z', fontsize=10)

# Enable subplot grid lines
ax1.grid(True, linewidth=0.5, linestyle=':')
ax2.grid(True, linewidth=0.5, linestyle=':')   
ax3.grid(True, linewidth=0.5, linestyle=':')

Finally, we'll display the sub-plots on the figure by calling ax.plot and specifying the x-axis and y-axis deques. We'll also specify different colors for each graphs to help us visually identify the sub-graphs.

# Display the sub-plots
ax1.plot(x_vals, accel_x, color='r')
ax2.plot(x_vals, accel_y, color='g')
ax3.plot(x_vals, accel_z, color='b')

# Display the sub-plots
ax1.plot(x_vals, accel_x, color='r')
ax2.plot(x_vals, accel_y, color='g')
ax3.plot(x_vals, accel_z, color='b')

Finally, we'll pause the plot's drawing for INTERVAL seconds. 

# Pause the plot for INTERVAL seconds 
plt.pause(INTERVAL)

# Pause the plot for INTERVAL seconds 
plt.pause(INTERVAL)

This method "makes an animation by repeatedly calling a function", animate. We provide it the figure we generated earlier and the function we'd like to animate.

# Update graph every 125ms
ani = FuncAnimation(fig, animate)

# Update graph every 125ms
ani = FuncAnimation(fig, animate)

Wiring

We'll be using the MLX90640 IR Thermal Camera Breakout to display an animated image comprised of thermal data to our notebook.

This breakout contains a 24x32 array of IR thermal sensors. When it's connected to the MCP2221, it returns an array of 768 individual infrared temperature readings over I2C. We'll read these values, manipulate, and display them in our Jupyter Notebook.

The MCP2221 and MLX90640 both have STEMMA QT connectors, so you can either wire it up on a breadboard or use a STEMMA QT cable.

Then, download the example Jupyter notebook:

Code Usage

In the Jupyter file browser, click Upload. From the file browser, select the MLX90640 Thermal Camera.ipynb example.

Click run on the next cell. This cell reads data from the thermal camera and splits it into a 32x24 array of thermal readings. We're using the numpy package so we can do perform fast manipulations to the data within the array. 

The next cell plots the data read from the previous cell. Click run to see a heatmap of your data.

This sensor reads the data twice per frame, in a checker-board pattern, so it's normal to see a checker-board dither effect when moving the sensor around - the effect isn't noticable when things move slowly.

The final cell in this notebook produces a live heat-map from your thermal camera. Point the camera towards yourself and click run.

Note: This GIF has been sped up (2x), we suggest moving very slowly to avoid a dithering effect.