{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Основы анализа данных в Python\n",
"\n",
"*Тамбовцева Алла, НИУ ВШЭ*\n",
"\n",
"## Линейная регрессия с `statsmodels` и `seaborn`\n",
"\n",
"### Подготовка и описание данных"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Установим библиотеку `seaborn` для более продвинутой и при этом более удобной (по сравнению с `matplotlib`) визуализации данных:"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"!pip install seaborn"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Оператор `!` дает понять Jupyter, что выше мы написали не код на языке Python, а команду для установки, которая обычно пишется в консоли.\n",
"\n",
"*Примечание.* Если некоторые строки кода ниже, в частности, `set_theme()` не работают, это, скорее всего, из-за того, что на компьютере уже установлена более старая версия `seaborn`, в которой некоторые функции отсутствовали или выглядели иначе. Поэтому при запуске `!pip install seaborn` новая версия не была установлена. Чтобы это исправить, можно установить новую версию явно, написав `!pip install seaborn==0.12.0`."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Импортируем эту библиотеку с сокращенным названием `sns`, так часто делают в официальных тьюториалах:"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import seaborn as sns"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Заодно импортируем остальные библиотеки и модули для работы: `pandas` для работы с датафреймом и `pyplot` для базовых графиков:"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"from matplotlib import pyplot as plt"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Загрузим данные из CSV-файла по ссылке (можно сохранить файл и загрузить его локально, просто по названию):"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"df = pd.read_csv(\"https://vincentarelbundock.github.io/Rdatasets/csv/carData/Salaries.csv\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Посмотрим на первые несколько строк датафрейма:"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
Unnamed: 0
\n",
"
rank
\n",
"
discipline
\n",
"
yrs.since.phd
\n",
"
yrs.service
\n",
"
sex
\n",
"
salary
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
1
\n",
"
Prof
\n",
"
B
\n",
"
19
\n",
"
18
\n",
"
Male
\n",
"
139750
\n",
"
\n",
"
\n",
"
1
\n",
"
2
\n",
"
Prof
\n",
"
B
\n",
"
20
\n",
"
16
\n",
"
Male
\n",
"
173200
\n",
"
\n",
"
\n",
"
2
\n",
"
3
\n",
"
AsstProf
\n",
"
B
\n",
"
4
\n",
"
3
\n",
"
Male
\n",
"
79750
\n",
"
\n",
"
\n",
"
3
\n",
"
4
\n",
"
Prof
\n",
"
B
\n",
"
45
\n",
"
39
\n",
"
Male
\n",
"
115000
\n",
"
\n",
"
\n",
"
4
\n",
"
5
\n",
"
Prof
\n",
"
B
\n",
"
40
\n",
"
41
\n",
"
Male
\n",
"
141500
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Unnamed: 0 rank discipline yrs.since.phd yrs.service sex salary\n",
"0 1 Prof B 19 18 Male 139750\n",
"1 2 Prof B 20 16 Male 173200\n",
"2 3 AsstProf B 4 3 Male 79750\n",
"3 4 Prof B 45 39 Male 115000\n",
"4 5 Prof B 40 41 Male 141500"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"В этом файле сохранены данные по сотрудникам университета в США, а именно следующие их характеристики:\n",
"\n",
"* `rank`: должность;\n",
"* `discipline`: тип преподаваемой дисциплины (`A` – теоретическая, `B` – практическая);\n",
"* `yrs.since.phd`: число лет с момента получения степени PhD;\n",
"* `yrs.service`: число лет опыта работы;\n",
"* `sex`: пол;\n",
"* `salary`: заработная плата за 9 месяцев, в долларах.\n",
"\n",
"Для удобства дальнейшей работы давайте переименуем столбцы с длинными названиями: вместо `yrs.since.phd` запишем просто `phd`, а вместо `yrs.service` — `service`. Для этого нам понадобится метод `.rename()` для датафреймов:"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"df.rename(columns = {\"yrs.since.phd\" : \"phd\", \"yrs.service\" : \"service\"}, inplace = True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Метод `.rename()` по умолчанию работает не со столбцами, а со строками, поэтому, чтобы переименование применялось именно к столбцам, мы обязательно указываем название аргумента `columns`, внутри которого помещаем словарь с парами *старое название-новое название*. Метод `.rename()`, как и ряд других методов на датафреймах, возвращает измененную копию таблицы, но не вносит изменения в сам датафрейм. Поэтому, чтобы сохранить изменения, мы дописываем аргумент `inplace = True`, который заменяет названия прямо в исходном датафрейме. В общем, строка выше эквивалентна обычному присваиванию через `=` (только более элегантна):\n",
"\n",
" df = df.rename(columns = {\"yrs.since.phd\" : \"phd\", \"yrs.service\" : \"service\"})"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Прежде чем приступить к построению регрессионных моделей, которые покажут нам взаимосвязь между заработной платой сотрудников и их характеристиками, давайте посмотрим на их распределение. Построим гистограмму для заработной платы, для столбца `salary`:"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# несильно беспокоимся о настройках графика, добавим тол\n",
"\n",
"plt.hist(df[\"salary\"], color = \"cornflowerblue\", edgecolor = \"white\");"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Проинтерпретируем полученный график. Распределение заработной платы не является симметричным, оно немного скошено вправо – в выборке есть несколько сотрудников с нетипично высокой заработной платой (200 тысяч долларов и выше). Больше всего сотрудников с заработной платой около 100 тысяч долларов. Небольшую скошенность распределения можно заметить и по описательным статистикам:"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"count 397.000000\n",
"mean 113706.458438\n",
"std 30289.038695\n",
"min 57800.000000\n",
"25% 91000.000000\n",
"50% 107300.000000\n",
"75% 134185.000000\n",
"max 231545.000000\n",
"Name: salary, dtype: float64"
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df[\"salary\"].describe()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Среднее выборки равно 113706 долларов, а медиана чуть меньше, 107300 долларов. Среднее здесь немного «завышено», как раз из-за наличия сотрудников с нетипично высокой заработной платой. Отсюда и скошенность – распределение симметрично, если среднее и медиана совпадают. \n",
"\n",
"Теперь построим гистограмму для стажа, числа лет опыта работы:"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.hist(df[\"service\"], color = \"cornflowerblue\", edgecolor = \"white\");"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Тоже имеет место скошенность вправо, при этом большая часть значений сосредоточена в интервале от 0 до 10 (примерно). То есть, довольно много сотрудников с относительно небольшим опытом работы, не более 10 лет."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Изучение взаимосвязи: диаграммы рассеивания и корреляция"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Приступим к изучению взаимосвязи между опытом работы и заработной платой. Начнем с диаграммы рассеивания, только построим ее не с помощью `matplotlib`, а с помощью `seaborn`. Сначала настроим общую тему графиков – чтобы на них присутствовала сетка, а фон при этом был белым:"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"sns.set_theme(style = \"whitegrid\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Настройки сохранены, они будут применены ко всем типам графиков далее. Полный список настроек можно посмотреть в [документации](https://seaborn.pydata.org/generated/seaborn.set_theme.html) функции `set_theme()`. Теперь построим диаграмму рассеивания с помощью функции `scatterplot()`:"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"sns.scatterplot(data = df, x = \"service\", y = \"salary\");"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Устройство этой функции простое: в `data` указываем название датафрейма, столбцы из которого мы используем для построения графика, а затем в `x` и `y` перечисляем сами названия этих столбцов.\n",
"\n",
"Что мы видим на диаграмме рассеивания? Связь между числом лет опыта работы и заработной платы есть, но не то чтобы линейная. В целом, если стаж сотрудников меньше 30 лет, связь между опытом работы и заработной платы прямая (чем больше опыта, тем выше зарплата), а если стаж выше – связь обратная (чем больше опыта, тем ниже зарплата). Это вполне объяснимо: опыт работы естественным образом связан с возрастом, а с определенного возраста заработная плата может становиться ниже в силу разных причин (меньшая активность, неготовность обучаться новым технологиям, состояние здоровья и прочее).\n",
"\n",
"Действительно, если мы вычислим коэффициент корреляции Пирсона между этими показателями, который показывает линейную связь, он будет невысоким:"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
salary
\n",
"
service
\n",
"
\n",
" \n",
" \n",
"
\n",
"
salary
\n",
"
1.000000
\n",
"
0.334745
\n",
"
\n",
"
\n",
"
service
\n",
"
0.334745
\n",
"
1.000000
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" salary service\n",
"salary 1.000000 0.334745\n",
"service 0.334745 1.000000"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df[[\"salary\", \"service\"]].corr() # 0.33"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Можем дополнительно сгруппировать точки в зависимости от должности (`rank`) и учесть это на графике:"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"