Predicting-condominium-price-using-data-from-webscraping/003_data_cleaning_pred_current_price.py at master · ekapope/Predicting-condominium-price-using-data-from-webscraping · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
# -*- coding: utf-8 -*-
"""
This script will:
1. load df_completed.csv output from '002_Scrape_info_for_each_condo.py'
2. check NAs and data types for each column
3. perform data manipulation
    - clean each column using regex
    - change numbers from strings to numeric
    - impute missing values
4. export dataframe to 'df_cleaned_for_ML_regression.csv'

"""

# import packages
import re
from datetime import datetime
import time
import pandas as pd
import pickle as pk
from bs4 import BeautifulSoup
import matplotlib.pyplot as plt
import numpy as np
import requests
import os
os.chdir(r"D:\GitHub_Personal\2019-01-Web-Scraping-using-selenium-and-bs4")

#load csv
df= pd.read_csv("data\scraped_data\df_completed.csv", sep=',',encoding='utf-8-sig')

df.isnull().sum()
df.info()

#add id column
df['id'] = df.index

#clean up each column
#strip str
df['name'] = df['name'].str.strip()
df['district'] = df['district'].str.strip()

#replace " and change to numeric
df['latitude'] = pd.to_numeric(df['latitude'].str.replace('\"',''))
df['longitude'] = pd.to_numeric(df['longitude'].str.replace('\"',''))

#add building age 2020-'year_built'
df['bld_age'] = 2019-df['year_built']

#replace , and change to numeric
df['proj_area'] = pd.to_numeric(df['proj_area'].str.replace('\,',''))

#check no numeric rows, convert to numeric, replace with Nan
print(df[~df['units'].str.isnumeric()]['units'])
df['units'] = pd.to_numeric(df['units'], errors = 'coerce')
missing_idx = np.isnan(df['units'])
missing_unit_df = df[missing_idx]
plt.hist(df['units'][~missing_idx])
plt.show()
print(df['units'][~missing_idx].describe())
#fill na with median in the same district
df['units'] = df.groupby("district")["units"].transform(lambda x: x.fillna(x.median()))

def find_dist(input_str):
    input_str = str(input_str)
    dist = re.findall(r"[-+]?\d*\.\d+|[-+]?\d+", input_str)[0]
    unit = re.findall(r" km ", input_str)
    if (len(unit)!=0): dist_km = float(dist)
    else: dist_km = float(dist)/1000
    return(dist_km)

#check shop col
df['shops']=df['shops'].str.replace('\'','').str.split('\,')
df['shops'][0]
col_list = ['dist_shop_'+str(i) for i in range(1, 6)]
# expand list into its own dataframe
df[col_list]= df['shops'].apply(pd.Series)
# loop all columns, result in distance (km)
for col in col_list: df[col]=df[col].apply(lambda x: find_dist(x))

#check schools col
df['schools']=df['schools'].str.replace('\"','\'').str.split('\', \'')
df['schools'][0]
len_school = df['schools'].apply(lambda x: len(x))
# expand list into its own dataframe
col_list = ['dist_school_'+str(i) for i in range(1, 6)]
df[col_list]= df['schools'].apply(pd.Series)
# loop all columns, result in distance (km)
for col in col_list: df[col]=df[col].apply(lambda x: find_dist(x))

#check restaurants col
df['restaurants']=df['restaurants'].str.replace('\"','\'').str.split('\', \'')
df['restaurants'][0]
col_list = ['dist_food_'+str(i) for i in range(1, 6)]
# expand list into its own dataframe
df[col_list]= df['restaurants'].apply(pd.Series)
# loop all columns, result in distance (km)
for col in col_list: df[col]=df[col].apply(lambda x: find_dist(x))

#check hospital col
df['hospital'][0]
df['hospital']=df['hospital'].apply(lambda x: find_dist(x))

#amenities col
#split into columns
#Elevator,Parking,Security,CCTV,Pool,Sauna,Gym,Garden,Playground,Shop,Restaurant,Wifi
col_list = ['Elevator','Parking','Security','CCTV','Pool','Sauna','Gym',\
            'Garden','Playground','Shop','Restaurant','Wifi']
df['amenities']=df['amenities'].str.replace('\'','')
len(df['amenities'][0])
df['amenities'][0]
df[col_list] = df['amenities'].str.split(",",expand=True,)
df[col_list] = df[col_list].apply(lambda x: x.str.strip())
df[col_list] = df[col_list].apply(lambda x: x.str.replace('\[','')).\
apply(lambda x: x.str.replace('\]',''))
df[col_list] = df[col_list].apply(pd.to_numeric)

#price_sqm
plt.hist(df['price_sqm'])
plt.show()
print(df['price_sqm'].describe())

# check transportation col
df['transportation']=df['transportation'].str.replace('\'','').str.split('\,')
# check the split
len_chk = df['transportation'].apply(lambda x: len(x))
df['transportation'][0]

# element number 0,3,6,9,12 are station types
col_list = ['tran_type'+str(i) for i in range(1, 6)]
df[col_list]= df['transportation'].apply(pd.Series).iloc[:,[0,3,6,9,12]]
for col in col_list: df[col]=df[col].apply(lambda x: re.findall("(expressway|mrt|bts)", x)[0])
# element number 1,4,7,10,13 are station names
col_list = ['tran_name'+str(i) for i in range(1, 6)]
df[col_list]= df['transportation'].apply(pd.Series).iloc[:,[1,4,7,10,13]].apply(lambda x: x.str.strip())
# element number 2,5,8,11,14 are distance to station
col_list = ['dist_tran_'+str(i) for i in range(1, 6)]
# expand list into its own dataframe
df[col_list]= df['transportation'].apply(pd.Series).iloc[:,[2,5,8,11,14]]
# loop all columns, result in distance (km)
for col in col_list: df[col]=df[col].apply(lambda x: find_dist(x))

#drop id, name, date columns
df.drop(['shops', 'schools', 'restaurants', 'amenities',
         'transportation','change_last_q', 'change_last_y', 'rental_yield',
         'change_last_y_rental_price', 'price_hist'], axis=1, inplace=True)

df.isnull().sum()
df.info()

#export to csv
df.to_csv(r"data\regression_data\df_cleaned_for_ML_regression.csv", index=False, encoding='utf-8-sig')