{
"cells": [
{
"cell_type": "markdown",
"id": "eff35ebd-8fa2-4b5d-8338-822e440c4289",
"metadata": {},
"source": [
"# Smooth Statistics Labels\n",
"\n",
"Smooth layers (e.g., `geom_smooth()`) now compute additional statistics that describe the fitted model quality and the fitted equation. These values are available as smooth stat variables (such as `..r2..` and `..adjr2..`) and can be displayed in a layer annotation via `smooth_labels`.\n",
"\n",
"`smooth_labels` is an annotation helper designed specifically for smooth layers. It extends `layer_labels`, so the familiar label-building API (`line()`, `format()`, `size()`, etc.) works the same way, while adding support for smooth-stat variables and the equation marker.\n",
"\n",
"Supported variables and markers:\n",
"\n",
"* `..r2..` — **R² (coefficient of determination)**. A goodness-of-fit measure showing what fraction of the variance in the response is explained by the fitted model. Values are typically between 0 and 1 (higher means the model explains more of the observed variation).\n",
"* `..adjr2..` — **adjusted R²**. A variant of R² that accounts for model complexity: it penalizes adding extra terms/parameters and is therefore more suitable for comparing models with different numbers of predictors (e.g., different polynomial degrees). Adjusted R² can be lower than R² and may even be negative for a very poor fit.\n",
"* `~eq` — **fitted equation**. Inserts the model equation into the annotation (can be configured with `eq()`)."
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "7a5c9a56-3edd-4109-ad8c-ef9ae1d1b64c",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
" \n",
" \n",
" "
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"import numpy as np\n",
"import pandas as pd\n",
"from lets_plot import *\n",
"LetsPlot.setup_html()"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "bbb9afae-832d-4bdf-8a12-363d2dd26462",
"metadata": {},
"outputs": [],
"source": [
"np.random.seed(42)"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "cc74a604-e64a-4e3c-890e-8e12d244886f",
"metadata": {},
"outputs": [],
"source": [
"t = np.linspace(0, 1, 100)\n",
"mean = 1 + np.zeros(2)\n",
"cov = np.eye(2)\n",
"x, y = np.random.multivariate_normal(mean, cov, t.size).T\n",
"df = pd.DataFrame({'t': t, 'x': x, 'y': y})\n",
"df = df.melt(id_vars=['t'], value_vars=['x', 'y'])"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "a3383095-3f77-4a29-8e0b-cbc0fc6c85ba",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"
\n",
" \n",
" \n",
" \n",
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"gggrid([\n",
" # Replaces 'y' with f(x) on the left-hand side of the equation\n",
" plot + geom_smooth(deg=2, labels=smooth_labels().line('~eq').eq(lhs='f(x)')),\n",
" # Completely removes the left-hand side of the equation\n",
" plot + geom_smooth(deg=2, labels=smooth_labels().line('~eq').eq(lhs='')),\n",
" # Replaces 'x' with 't' on the right-hand side of the equation\n",
" plot + geom_smooth(deg=2, labels=smooth_labels().line('~eq').eq(rhs='t')),\n",
"], ncol = 1)\n"
]
},
{
"cell_type": "markdown",
"id": "c3a6b410-02ed-4857-8ab6-4c82a0aa559c",
"metadata": {},
"source": [
"Coefficient formatting"
]
},
{
"cell_type": "code",
"execution_count": 14,
"id": "83945419-dd63-4819-9413-a58568d1a6e1",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"
\n",
"
\n",
" \n",
" \n",
"
\n",
" \n",
"
\n",
" \n",
" \n",
" \n",
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"gggrid([\n",
" # The regular format() is applied to all coefficients\n",
" plot + geom_smooth(deg=2, labels=smooth_labels().line('~eq').format('~eq', '.1f')),\n",
" # One format in eq() is applied to all coefficients\n",
" plot + geom_smooth(deg=2, labels=smooth_labels().line('~eq').eq(format='.2f')),\n",
" # You can pass a list to format each coefficient separately.\n",
" # The list length does not have to match the number of coefficients:\n",
" # the last format in the list will be used for any remaining coefficients.\n",
" plot + geom_smooth(deg=2, labels=smooth_labels().line('~eq').eq(format=['.1f', '.3f'])),\n",
"])"
]
},
{
"cell_type": "markdown",
"id": "7756da6f-e209-42d2-b305-48ffeca09214",
"metadata": {},
"source": [
"Parameter `threshold` "
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "3a7ca4e2-c485-4fbf-90ed-41b2d9d2c3c3",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"
\n",
"
\n",
" \n",
" \n",
"
\n",
" \n",
"
\n",
" \n",
" \n",
" \n",
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 15,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"plot + geom_smooth(deg=2, labels=smooth_labels()\n",
" .line('~eq')\n",
" .eq(threshold=0.01))"
]
},
{
"cell_type": "markdown",
"id": "816e6ff4-8d8b-486d-805c-a4fa761856fa",
"metadata": {},
"source": [
"#### Positioning: `label_x()` and `label_y()`\n",
"\n",
"By default, the text is shown in the upper-left corner. To change its position, use `label_x()` and `label_y()`. They accept strings (`left`, `center`, `right`) for `label_x()` and (`top`, `middle`, `bottom`) for `label_y()`, or an exact position in plot coordinates."
]
},
{
"cell_type": "code",
"execution_count": 16,
"id": "db06defc-f002-4761-ba1f-5b5cc6520585",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"
\n",
"
\n",
" \n",
" \n",
"
\n",
" \n",
"
\n",
" \n",
" \n",
" \n",
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 16,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"gggrid([\n",
" plot + geom_smooth(deg=2, labels = smooth_labels().label_x('center').label_y('middle')), \n",
" plot + geom_smooth(deg=2, labels = smooth_labels().label_x(0.5).label_y(4))\n",
"])"
]
},
{
"cell_type": "markdown",
"id": "d5b3a4d4-9df4-4457-8ad5-d9dd041f4025",
"metadata": {},
"source": [
"#### Grouping\n",
"If the data for `geom_smooth()` is grouped, the statistics are computed separately for each group automatically. If groups are differentiated by color, use `inherit_color()` to make the label text color match the geom_smooth() line color.\n",
"\n",
"By default, group labels are stacked vertically, but you can control the position of each group independently. To do this, pass lists to `label_x()` and `label_y()`.\n"
]
},
{
"cell_type": "code",
"execution_count": 17,
"id": "ca657b16-ec18-4100-a1b9-5396315f9c07",
"metadata": {},
"outputs": [],
"source": [
"x = [0, 1.5, 1.7, 2, 0, 1.5, 1.7, 2]\n",
"y = [0, 1, 1.8, 4, 0.5, 1.5, 3, 4.5]\n",
"g = ['a', 'a', 'a', 'a', 'b', 'b', 'b', 'b']\n",
"\n",
"plot_groups = ggplot({'x': x, 'y': y, 'g': g}, aes('x', 'y')) + geom_point() "
]
},
{
"cell_type": "code",
"execution_count": 18,
"id": "1306958c-1865-482c-8a28-4b45080eedb2",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"