`PyFixest` integrates with the excellent [marginaleffects](https://github.com/vincentarelbundock/pymarginaleffects) package.
Among many other things, `marginaleffects` allows you to compute:

- average marginal effects,
- subgroup marginal effects,
- linear and non-linear hypothesis tests.

```{python}
import numpy as np
from marginaleffects import avg_slopes, hypotheses

import pyfixest as pf

data = pf.get_data()
data["Y_bin"] = np.where(data["Y"] > 0, 1, 0)

fit = pf.feglm("Y_bin ~ X1 + X2 | f1", data=data, family="logit")
fit.tidy()
```

## Average marginal effects

To compute the average marginal effect of `X1` on the response scale:

```{python}
avg_slopes(fit, variables="X1")
```

You can also report group-specific average marginal effects:

```{python}
avg_slopes(fit, variables="X1", by="f1")
```

## Linear hypothesis tests

Suppose we want to test the linear restriction $X_1 = X_2$.
The `hypotheses()` function uses the model's estimated covariance matrix, so it works naturally with `PyFixest` objects.

```{python}
hypotheses(fit, "X1 - X2 = 0")
```

## Non-linear hypothesis tests

For non-linear transformations, `marginaleffects` applies the delta method automatically.
This is useful for ratio-style summaries or uplift calculations.

Here is a simple OLS example:

```{python}
fit_ols = pf.feols("Y ~ X1 + X2", data=data)
```

```{python}
hypotheses(fit_ols, "(X1 / Intercept - 1) * 100 = 0")
```

## Notes


For broader functionality / the full capability of `marginaleffects`, see the [marginaleffects book](https://marginaleffects.com/index.html).