[Python] Fitting sigmoid functions

I have some data I am trying to fit to a sigmoid function (several parameters for several species), however apparently some of them don't look fitted in a good way, I don't understand where the mistake is. I also got some warning:

RuntimeWarning: overflow encountered in exp
y = a / (1 + np.exp(-c * (x - d))) + b
OptimizeWarning: Covariance of the parameters could not be estimated
warnings.warn('Covariance of the parameters could not be estimated'
​

But I don't really know how to improve this

def sigmoid(x, a, b, c, d):
y = a / (1 + np.exp(-c * (x - d))) + b
return (y)

fig, axes = plt.subplots(
nrows=len(params),
ncols=len(species_list),
figsize=(15, 18),
sharex='col',
sharey='row'
)

legend_handles = []
legend_labels = []

for i, param in enumerate(params):
for j, species in enumerate(species_list):
ax = axes[i, j]
species_subset = df[df['Species'] == species]

for treatment in species_subset['Treatment'].unique():
treatment_subset = species_subset[species_subset['Treatment'] == treatment]

x = treatment_subset['RWC'].values
y = treatment_subset[param].values
mask = np.isfinite(x) & np.isfinite(y)
x_clean = x[mask]
y_clean = y[mask]

scatter = ax.scatter(x_clean, y_clean, label=treatment, alpha=0.6)

if len(x_clean) > 2:
try:
p0 = [max(y_clean), np.median(x_clean),1,min(y_clean)]
popt, _ = curve_fit(sigmoid, x_clean, y_clean, p0, maxfev=10000)
x_fit = np.linspace(x_clean.min(), x_clean.max(), 100)
y_fit = sigmoid(x_fit, *popt)
ax.plot(x_fit, y_fit, linestyle='--')
except RuntimeError:
print(f"Fit failed for {species}, {param}, {treatment}")
else:
print(f"Not enough data for {species}, {param}, {treatment} — skipping fit")

if i == 0 and j == 0:
legend_handles.append(scatter)
legend_labels.append(treatment)

if i == 0:
ax.set_title(species)
if j == 0:
ax.set_ylabel(param)
ax.set_xlabel('RWC')
ax.tick_params(axis='x', rotation=45)
ax.grid(True)

fig.legend(
legend_handles,
legend_labels,
title='Treatment',
loc='upper right',
bbox_to_anchor=(0.99, 0.99)
)

plt.tight_layout(rect=[0, 0, 0.95, 1])
plt.show()


This is the output I get:

Continue reading...