Do you know the automatic scientific notation form...

Below are several increasingly “aggressive” ways to force Matplotlib to stop switching your y-axis to scientific notation and instead show fixed-point labels such as 12.3, 0.4, etc. Pick the one that best matches how broadly you want to apply the rule.

---

1. Quick per-axis switch-off (good for most cases)

python
ax.ticklabel_format(axis='y',
                    style='plain',   # turn off exponential formatting
                    useOffset=False, # turn off the  +1.23e4 offset line
                    useMathText=False)

style='plain' tells the default ScalarFormatter to write every tick as a normal number, while useOffset=False prevents the offset line that sometimes appears above the axis. stackoverflow.com

---

2. Explicit fixed-point format (locks in the number of decimals)

python
from matplotlib.ticker import FormatStrFormatter

ax.yaxis.set_major_formatter(FormatStrFormatter('%.1f'))  # one decimal place

FormatStrFormatter bypasses the default formatter completely, so even very large or very small values are rendered with the string format you supply (%.1f, %.2f, {x:.3f}, etc.). stackoverflow.com

---

3. “Never ever use sci-notation” ScalarFormatter

python
from matplotlib.ticker import ScalarFormatter

fmt = ScalarFormatter()
fmt.set_scientific(False)          # don’t use 1.23e4 style
fmt.set_useOffset(False)           # don’t add +1e4 offset
fmt.set_powerlimits((-100, 100))   # never switch to *10^k   (very aggressive)
ax.yaxis.set_major_formatter(fmt)

set_powerlimits((-100, 100)) widens the threshold far beyond typical data ranges so the formatter will not flip to exponential form. matplotlib.org

---

4. Session-wide “scorched earth” setting (affects every plot you’ll make)

python
import matplotlib as mpl

mpl.rcParams['axes.formatter.useoffset'] = False
mpl.rcParams['axes.formatter.limits']    = (-100, 100)  # global version of powerlimits

Put those two lines once, early in your session (or in your matplotlibrc) and all subsequent figures will suppress scientific notation unless you explicitly override it. sqlpey.com

---

Tips & gotchas

Problem	Fix
Labels overlap after switching format	Rotate or nudge with plt.setp(ax.get_yticklabels(), rotation=0, ha='right') or use a MaxNLocator.
You only need this for one axis	Apply any approach to ax.xaxis or ax.yaxis individually.
Mixed log & linear plots	For a log axis use LogFormatter instead; the methods above apply to linear axes.

With these tools—especially option 3 or 4—you can completely eliminate scientific notation from your y-axis, no matter how extreme the underlying numbers.