{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Auto Miles Per Gallon Data Experiments\n",
"\n",
"This notebook runs the experiments and generates the figures on the [auto mpg data](https://archive.ics.uci.edu/ml/datasets/auto+mpg). We are working with a subset of the data created by: \n",
" \n",
" 1. selecting a subset of columns to suit our problem setting: three continuous (mpg, acceleration, and horsepower) and three categorical attributes (cylinders, model year, and origin) \n",
" 1. removing incomplete records\n",
" \n",
"This notebook generates the results and figures related to the AutoMPG experiments"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import numpy as np\n",
"import pandas as pd\n",
"import seaborn as sns\n",
"import matplotlib.colors as mcolors\n",
"import matplotlib.pyplot as plt\n",
"\n",
"# clean up notebook output by removing warnings about future changes\n",
"import warnings\n",
"warnings.simplefilter(action='ignore', category=FutureWarning)\n",
"\n",
"# our code packaged for easy use\n",
"import detect_simpsons_paradox as dsp"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
mpg
\n",
"
cylinders
\n",
"
horsepower
\n",
"
acceleration
\n",
"
model year
\n",
"
origin
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
18.0
\n",
"
8
\n",
"
130.0
\n",
"
12.0
\n",
"
70
\n",
"
1
\n",
"
\n",
"
\n",
"
1
\n",
"
15.0
\n",
"
8
\n",
"
165.0
\n",
"
11.5
\n",
"
70
\n",
"
1
\n",
"
\n",
"
\n",
"
2
\n",
"
18.0
\n",
"
8
\n",
"
150.0
\n",
"
11.0
\n",
"
70
\n",
"
1
\n",
"
\n",
"
\n",
"
3
\n",
"
16.0
\n",
"
8
\n",
"
150.0
\n",
"
12.0
\n",
"
70
\n",
"
1
\n",
"
\n",
"
\n",
"
4
\n",
"
17.0
\n",
"
8
\n",
"
140.0
\n",
"
10.5
\n",
"
70
\n",
"
1
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" mpg cylinders horsepower acceleration model year origin\n",
"0 18.0 8 130.0 12.0 70 1\n",
"1 15.0 8 165.0 11.5 70 1\n",
"2 18.0 8 150.0 11.0 70 1\n",
"3 16.0 8 150.0 12.0 70 1\n",
"4 17.0 8 140.0 10.5 70 1"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# import the prepared copy of the data\n",
"auto_df = pd.read_csv('../../../data/auto2.csv')\n",
"auto_df.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"From examining the above, we know that the integer columns are the group-by variables and the float type variables are the continuous attributes. The detector function will automatically "
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['cylinders', 'model year', 'origin']\n"
]
}
],
"source": [
"groupbyAttrs = auto_df.select_dtypes(include=['int64'])\n",
"groupbyAttrs_labels = list(groupbyAttrs)\n",
"print(groupbyAttrs_labels)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['mpg', 'horsepower', 'acceleration']\n"
]
}
],
"source": [
"continuousAttrs = auto_df.select_dtypes(include=['float64'])\n",
"continuousAttrs_labels = list(continuousAttrs)\n",
"print(continuousAttrs_labels)"
]
},
{
"cell_type": "markdown",
"metadata": {
"collapsed": true
},
"source": [
"## Results\n",
"\n",
"Now we can run our algorithm and print out the results after a little bit of post-processing to improve readability."
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
allCorr
\n",
"
attr1
\n",
"
attr2
\n",
"
reverseCorr
\n",
"
groupbyAttr
\n",
"
subgroup
\n",
"
\n",
" \n",
" \n",
"
\n",
"
1
\n",
"
0.423329
\n",
"
mpg
\n",
"
acceleration
\n",
"
-0.818873
\n",
"
cylinders
\n",
"
3
\n",
"
\n",
"
\n",
"
3
\n",
"
0.423329
\n",
"
mpg
\n",
"
acceleration
\n",
"
-0.341214
\n",
"
cylinders
\n",
"
6
\n",
"
\n",
"
\n",
"
4
\n",
"
0.423329
\n",
"
mpg
\n",
"
acceleration
\n",
"
-0.050545
\n",
"
model year
\n",
"
75
\n",
"
\n",
"
\n",
"
5
\n",
"
0.423329
\n",
"
mpg
\n",
"
acceleration
\n",
"
-0.051280
\n",
"
model year
\n",
"
79
\n",
"
\n",
"
\n",
"
0
\n",
"
-0.778427
\n",
"
mpg
\n",
"
horsepower
\n",
"
0.620807
\n",
"
cylinders
\n",
"
3
\n",
"
\n",
"
\n",
"
2
\n",
"
-0.778427
\n",
"
mpg
\n",
"
horsepower
\n",
"
0.013135
\n",
"
cylinders
\n",
"
6
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" allCorr attr1 attr2 reverseCorr groupbyAttr subgroup\n",
"1 0.423329 mpg acceleration -0.818873 cylinders 3\n",
"3 0.423329 mpg acceleration -0.341214 cylinders 6\n",
"4 0.423329 mpg acceleration -0.050545 model year 75\n",
"5 0.423329 mpg acceleration -0.051280 model year 79\n",
"0 -0.778427 mpg horsepower 0.620807 cylinders 3\n",
"2 -0.778427 mpg horsepower 0.013135 cylinders 6"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# run detection algorithm\n",
"result_df = dsp.detect_simpsons_paradox(auto_df)\n",
"\n",
"# Map attribute index to attribute name\n",
"result_df['attr1'] = result_df['attr1'].map(lambda x:continuousAttrs_labels[x])\n",
"result_df['attr2'] = result_df['attr2'].map(lambda x:continuousAttrs_labels[x])\n",
"# sort for easy reading\n",
"result_df = result_df.sort_values(['attr1', 'attr2'], ascending=[1, 1])\n",
"\n",
"# data frames print neatly in notebooks \n",
"result_df"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Plotting\n",
"\n",
"We plot all data in scatter plots based on each group by attribute, for each pair of candidate attributes. For each plot we add the overall trendline and the trend line for each occurence of Simpson's Paradox. "
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[8 4 6 3 5]\n",
"[70 71 72 73 74 75 76 77 78 79 80 81 82]\n"
]
}
],
"source": [
"print(auto_df.cylinders.unique())\n",
"print(auto_df['model year'].unique())"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"