fill_between: Shade Above and Below a Baseline

A line shows a trend; shading shows a verdict. That distinction is the whole point of fill_between(): it paints the area between two lines, and the where= argument masks that fill so "above the baseline" goes green and "below it" goes red in one chart. Color a row of numbers that way and the who-beat-average story reads at a glance. The same tool, pointed elsewhere, gives you confidence bands, projection ranges, and highlighted regions — but the green/red verdict is where I'd start.

This builds on Your First Visualization. The data is the bundled sample_standings.csv (real 2023 MLB standings), so it runs offline.

1. Set up a line and a baseline

   Sort the teams by run differential so the line sweeps low to high, then compute the baseline we'll shade against — the league mean, which for run differential is essentially zero.

   import numpy as np
   import pandas as pd
   
   df = pd.read_csv("sample_standings.csv").sort_values("RunDiff").reset_index(drop=True)
   x = np.arange(len(df))
   rd = df["RunDiff"].values
   mean = rd.mean()
   print("mean:", round(mean, 1), "| teams above:", int((rd >= mean).sum()))

   The baseline and how many clear it

   League mean run differential: 0.0  (≈0, as it must be)
   Teams above the mean: 13 of 30

   With the data sorted, every team left of the crossover sits below average and every team right of it sits above — the shape fill_between will color in.

2. Two fills, split by a where mask

   Call fill_between twice with the same line and baseline but opposite boolean masks. interpolate=True is the detail that matters: it closes each shaded wedge exactly where the line crosses the baseline, instead of leaving a ragged gap at the last data point.

   import matplotlib.pyplot as plt
   
   fig, ax = plt.subplots(figsize=(9, 5.5))
   ax.plot(x, rd, color="#20242B", lw=1.5)
   ax.axhline(mean, color="#6C7079", lw=1, ls="--")
   ax.fill_between(x, rd, mean, where=(rd >= mean), interpolate=True,
                   color="#2E7D4F", alpha=0.55, label="above average")
   ax.fill_between(x, rd, mean, where=(rd < mean), interpolate=True,
                   color="#B23A3A", alpha=0.55, label="below average")
   ax.set_xticks(x); ax.set_xticklabels(df["Abbr"], rotation=90)
   ax.legend()
   fig.savefig("fill_between.png", dpi=144, bbox_inches="tight")

   

   The eye reads the split instantly: a red basin of below-average teams on the left rising through the crossover into a green slope of above-average teams on the right. One chart, one glance, the whole league sorted into beat-average and didn't.

3. The same tool draws a band

   fill_between isn't just for baselines. Give it a lower and an upper series instead of a line and a constant, and it shades the band between them — the standard way to draw a ±1 standard-deviation range, a min-max envelope, or a projection's uncertainty around a rolling average.

   # Sketch: a shaded ±1σ band around a baseline
   sd = rd.std()
   ax.fill_between(x, mean - sd, mean + sd, color="#9A8F79", alpha=0.2)  # the "typical" zone

   That one idea — the area between two curves — covers a huge share of the shaded charts you see in analytics: error bands on a forecast, the gap between expected and actual, or simply the region a value spends most of its time in.

Troubleshooting