{
"cells": [
{
"cell_type": "code",
"id": "initial_id",
"metadata": {
"collapsed": true,
"ExecuteTime": {
"end_time": "2025-07-08T15:58:03.741302Z",
"start_time": "2025-07-08T15:58:03.737218Z"
}
},
"source": [
"import pandas as pd\n",
"import matplotlib.pyplot as plt\n",
"import seaborn as sns\n",
"\n",
"from misc import load_csv_dataset"
],
"outputs": [],
"execution_count": 1
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-07-08T15:58:27.047640Z",
"start_time": "2025-07-08T15:58:06.842257Z"
}
},
"cell_type": "code",
"source": [
"data = load_csv_dataset(\"names_featured.csv\")\n",
"df = pd.DataFrame(data)"
],
"id": "74a8a262bc55d976",
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
">> Loading CSV dataset from names_featured.csv\n",
">> Detected fieldnames: ['name', 'region', 'year', 'words', 'length', 'probable_native', 'probable_surname', 'sex']\n",
">> Successfully loaded with UTF-8 encoding\n"
]
}
],
"execution_count": 2
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-07-08T15:58:50.742471Z",
"start_time": "2025-07-08T15:58:45.080832Z"
}
},
"cell_type": "code",
"source": [
"df['year'] = pd.to_numeric(df['year'], errors='coerce')\n",
"df['words'] = pd.to_numeric(df['words'], errors='coerce')\n",
"df['length'] = pd.to_numeric(df['length'], errors='coerce')"
],
"id": "ebb18cabd40b011f",
"outputs": [],
"execution_count": 3
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-07-08T15:59:02.783458Z",
"start_time": "2025-07-08T15:59:02.297971Z"
}
},
"cell_type": "code",
"source": "df.describe().T",
"id": "295eda175081129c",
"outputs": [
{
"data": {
"text/plain": [
" count mean std min 25% 50% 75% \\\n",
"year 4418337.0 2017.038495 4.656899 2008.0 2013.0 2018.0 2021.0 \n",
"words 4418337.0 2.862385 0.473601 1.0 3.0 3.0 3.0 \n",
"length 4418337.0 18.216613 3.498679 1.0 16.0 18.0 21.0 \n",
"\n",
" max \n",
"year 2023.0 \n",
"words 10.0 \n",
"length 49.0 "
],
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" count | \n",
" mean | \n",
" std | \n",
" min | \n",
" 25% | \n",
" 50% | \n",
" 75% | \n",
" max | \n",
"
\n",
" \n",
" \n",
" \n",
" | year | \n",
" 4418337.0 | \n",
" 2017.038495 | \n",
" 4.656899 | \n",
" 2008.0 | \n",
" 2013.0 | \n",
" 2018.0 | \n",
" 2021.0 | \n",
" 2023.0 | \n",
"
\n",
" \n",
" | words | \n",
" 4418337.0 | \n",
" 2.862385 | \n",
" 0.473601 | \n",
" 1.0 | \n",
" 3.0 | \n",
" 3.0 | \n",
" 3.0 | \n",
" 10.0 | \n",
"
\n",
" \n",
" | length | \n",
" 4418337.0 | \n",
" 18.216613 | \n",
" 3.498679 | \n",
" 1.0 | \n",
" 16.0 | \n",
" 18.0 | \n",
" 21.0 | \n",
" 49.0 | \n",
"
\n",
" \n",
"
\n",
"