Your First Sports Data Visualization with matplotlib

By the end of this page you'll have a clean, labeled bar chart of the highest-scoring offenses of 2023, saved as an image you'd actually be willing to put in front of someone. But the chart is almost beside the point. What I really want you to walk away with is matplotlib's figure and axes model - because once that clicks, every other chart on this site is a variation on the same three or four lines, and you stop copy-pasting plotting code you don't understand.

Most beginners (me, a few years ago) learn matplotlib by collecting magic incantations off Stack Overflow and never quite knowing why they work. We're going to do the opposite: learn the small mental model first, then the chart almost draws itself. We'll use the same bundled sample_standings.csv from the last tutorial - the 2023 MLB regular-season standings I pulled from the MLB Stats API in June 2026 - so you should already be comfortable loading a CSV and picking columns from the pandas tutorial.

1. Load and shape the data

   We load the standings, then prepare exactly the slice we want to draw: the 15 top-scoring teams. There's a small trick in the sort. We sort descending to find the top 15, then sort that subset ascending again, because a horizontal bar chart draws its first row at the bottom — sorting ascending makes the longest bar land on top where the eye expects it.

   import pandas as pd
   
   df = pd.read_csv("sample_standings.csv")
   top = df.sort_values("RS", ascending=False).head(15).sort_values("RS")
   print(top[["Team", "RS", "RA", "RunDiff"]].tail(5).to_string())

   The five highest-scoring teams

         Team   RS   RA  RunDiff
   5   Astros  827  698      129
   3     Rays  860  665      195
   7  Rangers  881  716      165
   2  Dodgers  906  699      207
   0   Braves  947  716      231

   Because top is sorted ascending, its last five rows are the biggest offenses — which is why we print .tail(5) here. The Braves' 947 runs sit at the very bottom of this printout but will sit at the very top of the chart. Always look at the data you're about to plot; it's the cheapest way to catch a mistake before it becomes a misleading picture.

2. Understand the figure/axes model

   This is the one concept worth slowing down for. plt.subplots() returns two things:

   • a figure (fig) — the whole canvas, including margins, title area, and the file you eventually save;
   • an axes (ax) — the actual plotting rectangle with its x- and y-axis, where the bars, labels, and ticks live.

   You draw onto the axes and you save the figure. Almost every method we call — ax.barh, ax.set_title, ax.set_xlabel — is a message to the axes. Keeping that distinction straight is what stops chart code from feeling like guesswork.

   import matplotlib.pyplot as plt
   
   fig, ax = plt.subplots(figsize=(8, 5))

   The figsize is in inches (width, height). At 8 by 5 we get a landscape rectangle with room for fifteen team labels stacked down the side without crowding.

3. Draw the bars and label everything

   Now we draw. ax.barh makes a horizontal bar chart — the right choice when your categories are text labels like team names, because horizontal labels are far easier to read than rotated vertical ones. We pass the team names for the y-axis and runs scored for the bar lengths.

   bars = ax.barh(top["Team"], top["RS"], color="#B23A3A")
   ax.set_title("Runs scored, 2023 (top 15 offenses)")
   ax.set_xlabel("runs scored")
   ax.set_ylabel("")
   ax.bar_label(bars, fmt="%d", padding=4, fontsize=9)
   ax.margins(x=0.08)

   Three touches turn a rough plot into a finished one. A clear title and x-axis label tell the reader what they're looking at without a caption. ax.bar_label prints each bar's exact value at its end — the fmt="%d" formats them as whole numbers — so readers get both the visual comparison and the precise figure. And ax.margins(x=0.08) adds a little breathing room on the right so those value labels don't collide with the edge of the plot. We deliberately blank the y-axis label with set_ylabel("") because the team names already make it obvious.

4. Save it at publication resolution

   A chart you can't share isn't finished. We save the figure — remember, the whole canvas — to a PNG. Two arguments matter most: dpi=144 gives a crisp image that holds up on high-resolution screens, and bbox_inches="tight" trims excess whitespace so the saved file is snug around the chart.

   fig.savefig("runs_per_team.png", dpi=144, bbox_inches="tight")

   

   There it is — the Braves' 947 runs lead at the top, exactly as the data preview promised, with every bar carrying its precise value. Notice we never called plt.show(); saving straight to a file is how charts are produced on this site, and it's also how you'd generate images in a script or report without a window ever popping up. Open the PNG and admire your first real visualization.

Troubleshooting