interpolate¶

In [1]:

site_url = "http://www.kyoritsu-pub.co.jp/bookdetail/9784320123656"
zip_url = "http://www.kyoritsu-pub.co.jp/app/file/goods_contents/2377.zip"
pdf_url = "http://www.kyoritsu-pub.co.jp/app/file/goods_contents/2324.pdf"

In [2]:

import numpy as np
import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
import seaborn as sns

%matplotlib inline

In [3]:

import requests
import io
import shutil
import zipfile
import os

In [4]:

data_dir = "suppl_20151001/data/missdata"
filename = "employee_IQ_JP.csv"
file_path = data_dir + "/" + filename
if not os.path.exists(file_path):
    # headers = {'user-agent': 'my-agent'}
    # r = requests.get(pdf_url, headers=headers)
    # shutil.copyfileobj(io.BytesIO(r.content), open(pdf_url.split("/")[-1], "wb"))

    proxies={
        "http": "http://xxx.xxx.xxx.:xxxx",
        "https": "http://xxx.xxx.xxx.:xxxx",
    }
    r = requests.get(zip_url, proxies=proxies)
    with zipfile.ZipFile(io.BytesIO(r.content)) as z:
        z.extractall()

df = pd.read_csv(file_path)
df.describe()

Out[4]:

	IQ	JobPerformance	MCAR	MCAR.is.missing	MAR	MAR.is.missing	MNAR	MNAR.is.missing
count	20.000000	20.000000	16.00000	20.000000	15.000000	20.000000	15.00000	20.000000
mean	100.000000	10.350000	10.56250	0.200000	10.666667	0.250000	11.40000	0.250000
std	14.127241	2.680829	2.82769	0.410391	2.794553	0.444262	2.22967	0.444262
min	78.000000	7.000000	7.00000	0.000000	7.000000	0.000000	9.00000	0.000000
25%	90.000000	8.750000	8.75000	0.000000	9.000000	0.000000	10.00000	0.000000
50%	97.500000	10.000000	10.00000	0.000000	10.000000	0.000000	11.00000	0.000000
75%	109.000000	12.000000	12.25000	0.000000	12.000000	0.250000	12.50000	0.250000
max	134.000000	16.000000	16.00000	1.000000	16.000000	1.000000	16.00000	1.000000

In [ ]:

# http://ejje.weblio.jp/content/interpolate

In [52]:

s = pd.Series([0, 1, np.nan, 3])
# help(s.interpolate)

In [40]:

methods = ['linear', 'time', 'index', 'values', 'nearest', 'zero',
              'slinear', 'quadratic', 'cubic', 'barycentric', 'krogh',
              'polynomial', 'spline', 'piecewise_polynomial', 'pchip']
methods.remove("time")
methods.remove("quadratic")
methods.remove("cubic")
methods.remove("polynomial")
methods.remove("spline")
methods.remove("piecewise_polynomial")

methods

Out[40]:

['linear',
 'index',
 'values',
 'nearest',
 'zero',
 'slinear',
 'barycentric',
 'krogh',
 'pchip']

In [51]:

def _interpolate(m):
    _s = s.interpolate(method=m)
    _s.name = m
    return _s
pd.concat(map(_interpolate, methods), axis=1)

Out[51]:

	linear	index	values	nearest	zero	slinear	barycentric	krogh	pchip
0	0	0	0	0	0	0	0	0	0
1	1	1	1	1	1	1	1	1	1
2	2	2	2	1	1	2	2	2	2
3	3	3	3	3	3	3	3	3	3