{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Stratified sampling\n",
"\n",
"In large dataset a relatively small group of points might be overplotted by the dominant group. In this case **stratified** sampling can help."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"execution": {
"iopub.execute_input": "2024-04-17T07:36:19.442338Z",
"iopub.status.busy": "2024-04-17T07:36:19.441932Z",
"iopub.status.idle": "2024-04-17T07:36:19.765711Z",
"shell.execute_reply": "2024-04-17T07:36:19.765421Z"
}
},
"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",
"\n",
"LetsPlot.setup_html()"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"execution": {
"iopub.execute_input": "2024-04-17T07:36:19.779642Z",
"iopub.status.busy": "2024-04-17T07:36:19.779377Z",
"iopub.status.idle": "2024-04-17T07:36:19.781708Z",
"shell.execute_reply": "2024-04-17T07:36:19.781479Z"
}
},
"outputs": [],
"source": [
"N = 5000 \n",
"small_group = 3\n",
"large_group = N - small_group\n",
"\n",
"np.random.seed(123)\n",
"data = dict(\n",
" x = np.random.normal(0, 1, N),\n",
" y = np.random.normal(0, 1, N),\n",
" cond = ['A' for _ in range(small_group)] + ['B' for _ in range(large_group)]\n",
")"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"execution": {
"iopub.execute_input": "2024-04-17T07:36:19.782743Z",
"iopub.status.busy": "2024-04-17T07:36:19.782665Z",
"iopub.status.idle": "2024-04-17T07:36:19.837983Z",
"shell.execute_reply": "2024-04-17T07:36:19.837776Z"
}
},
"outputs": [
{
"data": {
"text/html": [
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Data points in group 'A' (small group) are overplotted by the dominant group 'B'.\n",
"p = ggplot(data, aes('x','y',color='cond')) + \\\n",
" scale_color_manual(values=[\"red\", \"#1C9E77\"], breaks=['A', 'B'])\n",
"p + geom_point(size=5, alpha=.2)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"execution": {
"iopub.execute_input": "2024-04-17T07:36:19.839749Z",
"iopub.status.busy": "2024-04-17T07:36:19.839628Z",
"iopub.status.idle": "2024-04-17T07:36:19.851644Z",
"shell.execute_reply": "2024-04-17T07:36:19.851467Z"
}
},
"outputs": [
{
"data": {
"text/html": [
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# The 'random' sampling loses the group 'A' altogether.\n",
"p + geom_point(size=5, sampling=sampling_random(50, seed=2))"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"execution": {
"iopub.execute_input": "2024-04-17T07:36:19.852871Z",
"iopub.status.busy": "2024-04-17T07:36:19.852793Z",
"iopub.status.idle": "2024-04-17T07:36:19.864217Z",
"shell.execute_reply": "2024-04-17T07:36:19.864035Z"
}
},
"outputs": [
{
"data": {
"text/html": [
" \n",
" "
],
"text/plain": [
""
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Stratified sampling ensures that group 'A' is represented.\n",
"p + geom_point(size=5, sampling=sampling_random_stratified(50, seed=2))"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.13"
}
},
"nbformat": 4,
"nbformat_minor": 1
}