Matplotlib in Python - How to Plot Data

Data. The raw material of the digital people, an untouched block of marble. Yet, what is a data set if not visualized? Just as what is marble if not sculpted? Today we use Matplotlib as our chisel to carve out visual masterpieces from data. Time to turn numbers into images and noise into signal. Matplotlib is a library for creating static, animated, and interactive visualizations in Python. With Matplotlib, you can programmatically create two-dimensional plots. If you find yourself constrained by this planar existence, you may upgrade to three dimensions. And if a static snapshot feels insufficient, you can animate your plots, adding the fourth dimension of time. Enough plot talk. Time for plot action. Install Matplotlib. With our digital palette ready, let us create our first visualization. Import pyplot, abbreviated as plt. The plot method accepts a sequence of y values and the show will display the plot. Notice we did not specify x coordinates. In such cases, Matplotlib automatically assigns integer indices starting from zero. Plot. But perhaps you desire more control over where each point should be. Let us demonstrate with a parabola. To control the display of parabolic points, create two lists, one for x coordinates and another with the corresponding y values. When we pass both sequences to plot followed by show, we witness mathematics made visible. This is the dull default appearance of the plot. Yuck. Fortunately, Matplotlib provides extensive customization options to elevate your visualization. There are keyword arguments for specifying colors, data point markers, — and more. The title method adds context because even plots deserve proper introductions. Label your axes with xlabel and ylabel. Replot. Our humble parabola has an improved style game. For those of you who prefer your plots like you prefer your matter, dark and mysterious, the dark background style awaits your command. Remember to declare your styling intentions before plot creation. Plot again. Not satisfied with a single solitary plot? Let us explore data cohabitation. We shall plot a sine wave with amplitude 50 alongside its more volatile companion, the tangent function. Both shall share the domain of -5 to 5. We define our y values for each mathematical entity and graph them with separate plot commands. To prevent visual confusion, we assign distinct colors and labels. The legend method provides a legendary service. Plotting now. Notice a critical flaw in our implementation. The plot method creates a piecewise linear approximation. It connects the dots. Choppy and sloppy. To improve matters, increase the sampling frequency. That is, use more x coordinates that are closer together. Now we have a smoother, more accurate representation of these transcendental functions. Humans have a fascination with bars. So you will likely enjoy this. To create a bar chart, use the bar function. This transforms categorical data into rectangular representations of value. Speaking of value, allow me to share our recent merchandise analytics. Socratica offers high-quality Python-themed merchandise at everyday low prices. Here is our recent sales data for different Python products. Define your categories and corresponding values. Then call the bar function to create the chart. Do not forget to responsibly add a title and label the axes. Display. A masterpiece of retail data visualization. Available today on a website near you. In textbooks, probability distributions appear as elegant, smooth curves. But do not be fooled. In real life, data is choppy. So how do you visualize irregularly distributed data? With histograms. This technique groups data into bins and plots the frequency of each bin. To demonstrate, let us use NumPy to select 1,000 random numbers from a normal distribution. We will group this data into 30 bins and plot it with a hist function. Titles should be present and axes should be labeled. Show. Randomness has resolved into a recognizable pattern. In the search for meaning in the universe, one must not allow preconceptions to influence

observations. That brings us to scatter plots. The most unbiased form of data visualization. Scatter plots are a powerful tool for visualizing raw data, especially when determining if a relationship exists between two variables. Unlike the plot function full of assumptions, scatter plots do not connect the dots. They simply present each data point as it exists in mathematical space. Allow me to demonstrate with two uncorrelated variables. We will generate two sets of normally distributed random numbers. To create the plot, we will use the scatter function. We then sprinkle the plot with helpful text and show. There is no clear pattern or relationship between the points. The values of x and y are entirely uncorrelated. Which means knowing the value of one does not give us any information about the other. But sometimes the relationship truly does exist. And not purely as a figment of your fevered imagination. This time, we will generate data from a two-dimensional normal distribution with a correlation coefficient of 0. 7. If we use a scatter plot to examine this data, the relationship will be clear. Display. Data scattered, relationship revealed. In the cutthroat world of corporate finance, one question is on everyone's mind. Am I getting my fair share? There is a visualization for that. Pie charts. An efficient means of displaying proportions. Let us construct one to illustrate ownership shares in a company. Observe. Labels are defined. Corresponding sizes are allocated. The pie function will do the rest. One slice shall be exploded for emphasis. This technique is commonly used to prompt discussion. We assign colors for aesthetic differentiation. Blues and oranges. Utilitarian yet pleasing. Annotations are added. Percentages displayed. Show us this pie. Result. A visual representation of the corporate bigwig club. Proportional ownership conveyed. Line, histo, bar, pie, scatter. We have seen many a plot, but each as a singular experience. What if we wanted more? More dimensions, more perspectives, more plots? Behold the subplot. The perfect way to visualize multiple aspects of data in one clean, organized figure. First, we define our data. Simple values, a sine wave, some random points, a few categories, and a data set just waiting for a histogram. These are our building blocks. The magic begins with the subplots function. This creates a grid of four plots arranged two by two. The fig size ensures our figure is well-sized at 10 by 8 inches. Efficient. Not excessive. Now, let us unveil our four plots. In quadrant one, we place our sine wave, its oscillations rendered in blue. Adjacent, the scatter plot displays pure randomness. A definite contrast. Below, we revisit our old friend, the pie chart, showing proportions in all its circular glory. Finally, the bottom right holds a histogram, visualizing the frequency distribution of our data. Show plot display. And just like that, we see different data types, different stories, all in one space. A well-balanced feast for the analytical mind. But perhaps this visual arrangement deserves more than just a fleeting moment on screen. For posterity, we shall preserve this creation using savefig. After all, data visualization of this caliber deserves to be archived at 300 DPI. And now, let us embrace the unexpected. Matplotlib is not only for plotting data. It can also bring images to life. We begin by importing the image module. Next, we summon RoboKitty. The imshow function renders our digital feline. We then remove the axis with axis off. After all, a robotic cat requires no Cartesian plane to express its magnitude. We add a fitting title. The result? An enigmatic robotic feline enjoying a moment in the sun. Congratulations. You've completed your apprenticeship in the fine art of data visualization. No longer are you a mere data collector. You are a data artist, transforming raw numbers into visual narratives.

The next time you encounter a data set yearning to be visualized, you shall know exactly which tools to summon.