1925 lines
71 KiB
Text
1925 lines
71 KiB
Text
{
|
||
"cells": [
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 1,
|
||
"id": "f20bddee-1e7c-4c46-a17a-c7bb6c13f30c",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"/Users/kadir/code/realestate/crawler/venv/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
|
||
" from .autonotebook import tqdm as notebook_tqdm\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"from data_access import Listing\n",
|
||
"import pandas as pd"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 2,
|
||
"id": "b1101088-9613-465f-81fd-79801e0202b8",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"ls = Listing.get_all_listings()\n",
|
||
"\n",
|
||
"decisions = {\n",
|
||
" 142789514: 'n',\n",
|
||
" 136010102: 'n',\n",
|
||
" 141457334: 'y',\n",
|
||
" 86778015: 'n',\n",
|
||
" 134574563: 'n',\n",
|
||
" 86648925: 'n',\n",
|
||
" 143319068: 'n',\n",
|
||
" 135668207: 'n',\n",
|
||
" 142063949: 'n',\n",
|
||
" 145051769: 'n',\n",
|
||
" 138945719: 'n',\n",
|
||
" 135714833: 'n',\n",
|
||
" 144983192: 'n',\n",
|
||
" 144666920: 'n',\n",
|
||
" 143895080: 'n',\n",
|
||
" 141114200: 'n',\n",
|
||
" 145407389: 'n',\n",
|
||
" 145047533: 'n',\n",
|
||
" 145161722: 'n',\n",
|
||
" 145130066: 'n',\n",
|
||
" 142110470: 'n',\n",
|
||
" 133667606: 'n',\n",
|
||
" 145005536: 'n',\n",
|
||
" 143458961: 'n',\n",
|
||
" 141412010: 'y',\n",
|
||
" 138683339: 'n',\n",
|
||
" 138490370: 'n',\n",
|
||
" 137805509: 'n',\n",
|
||
" 135854261: 'n',\n",
|
||
"}"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 3,
|
||
"id": "63e61601-7e3f-4d58-89f6-1794e4868cc3",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"ds = [l.dict_nicely() for l in ls]"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 4,
|
||
"id": "1c222721-f426-42c0-9ac5-badc1f7a2034",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/html": [
|
||
"<div>\n",
|
||
"<style scoped>\n",
|
||
" .dataframe tbody tr th:only-of-type {\n",
|
||
" vertical-align: middle;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe tbody tr th {\n",
|
||
" vertical-align: top;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe thead th {\n",
|
||
" text-align: right;\n",
|
||
" }\n",
|
||
"</style>\n",
|
||
"<table border=\"1\" class=\"dataframe\">\n",
|
||
" <thead>\n",
|
||
" <tr style=\"text-align: right;\">\n",
|
||
" <th></th>\n",
|
||
" <th>identifier</th>\n",
|
||
" <th>sqm_ocr</th>\n",
|
||
" <th>price</th>\n",
|
||
" <th>price_per_sqm</th>\n",
|
||
" <th>url</th>\n",
|
||
" <th>bedrooms</th>\n",
|
||
" <th>travel_time_fastest</th>\n",
|
||
" <th>travel_time_second</th>\n",
|
||
" <th>lease_left</th>\n",
|
||
" <th>development</th>\n",
|
||
" <th>decision</th>\n",
|
||
" </tr>\n",
|
||
" </thead>\n",
|
||
" <tbody>\n",
|
||
" <tr>\n",
|
||
" <th>0</th>\n",
|
||
" <td>102360773</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>350000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/102360773</td>\n",
|
||
" <td>1</td>\n",
|
||
" <td>{'duration': 2695, 'distance': 6467, 'duration...</td>\n",
|
||
" <td>{'duration': 1682, 'distance': 6810, 'duration...</td>\n",
|
||
" <td>119.0</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>1</th>\n",
|
||
" <td>105836849</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>400000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/105836849</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 2565, 'distance': 14070, 'duratio...</td>\n",
|
||
" <td>{'duration': 2565, 'distance': 14070, 'duratio...</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>2</th>\n",
|
||
" <td>107233214</td>\n",
|
||
" <td>76.91</td>\n",
|
||
" <td>400000.0</td>\n",
|
||
" <td>5200.884150</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/107233214</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 1714, 'distance': 9570, 'duration...</td>\n",
|
||
" <td>{'duration': 1774, 'distance': 9570, 'duration...</td>\n",
|
||
" <td>91.0</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>3</th>\n",
|
||
" <td>107976896</td>\n",
|
||
" <td>112.40</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>7117.437722</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/107976896</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 1862, 'distance': 8278, 'duration...</td>\n",
|
||
" <td>{'duration': 1862, 'distance': 8278, 'duration...</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>4</th>\n",
|
||
" <td>115499441</td>\n",
|
||
" <td>115.60</td>\n",
|
||
" <td>775000.0</td>\n",
|
||
" <td>6704.152249</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/115499441</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 2943, 'distance': 7437, 'duration...</td>\n",
|
||
" <td>{'duration': 2167, 'distance': 9920, 'duration...</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>...</th>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" <td>...</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>2181</th>\n",
|
||
" <td>86813508</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86813508</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 2400, 'distance': 13983, 'duratio...</td>\n",
|
||
" <td>{'duration': 2605, 'distance': 14702, 'duratio...</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>True</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>2182</th>\n",
|
||
" <td>86813523</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>655000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86813523</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 2400, 'distance': 13983, 'duratio...</td>\n",
|
||
" <td>{'duration': 2605, 'distance': 14702, 'duratio...</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>True</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>2183</th>\n",
|
||
" <td>86814669</td>\n",
|
||
" <td>82.80</td>\n",
|
||
" <td>550000.0</td>\n",
|
||
" <td>6642.512077</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86814669</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 2310, 'distance': 12972, 'duratio...</td>\n",
|
||
" <td>{'duration': 2322, 'distance': 12104, 'duratio...</td>\n",
|
||
" <td>0.0</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>2184</th>\n",
|
||
" <td>86955958</td>\n",
|
||
" <td>90.00</td>\n",
|
||
" <td>300000.0</td>\n",
|
||
" <td>3333.333333</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86955958</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>{'duration': 2332, 'distance': 6898, 'duration...</td>\n",
|
||
" <td>{'duration': 2248, 'distance': 6893, 'duration...</td>\n",
|
||
" <td>105.0</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>2185</th>\n",
|
||
" <td>86972726</td>\n",
|
||
" <td>81.60</td>\n",
|
||
" <td>790000.0</td>\n",
|
||
" <td>9681.372549</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86972726</td>\n",
|
||
" <td>2</td>\n",
|
||
" <td>{'duration': 1393, 'distance': 6390, 'duration...</td>\n",
|
||
" <td>{'duration': 2733, 'distance': 6490, 'duration...</td>\n",
|
||
" <td>993.0</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" </tr>\n",
|
||
" </tbody>\n",
|
||
"</table>\n",
|
||
"<p>2186 rows × 11 columns</p>\n",
|
||
"</div>"
|
||
],
|
||
"text/plain": [
|
||
" identifier sqm_ocr price price_per_sqm \\\n",
|
||
"0 102360773 NaN 350000.0 NaN \n",
|
||
"1 105836849 NaN 400000.0 NaN \n",
|
||
"2 107233214 76.91 400000.0 5200.884150 \n",
|
||
"3 107976896 112.40 800000.0 7117.437722 \n",
|
||
"4 115499441 115.60 775000.0 6704.152249 \n",
|
||
"... ... ... ... ... \n",
|
||
"2181 86813508 NaN 750000.0 NaN \n",
|
||
"2182 86813523 NaN 655000.0 NaN \n",
|
||
"2183 86814669 82.80 550000.0 6642.512077 \n",
|
||
"2184 86955958 90.00 300000.0 3333.333333 \n",
|
||
"2185 86972726 81.60 790000.0 9681.372549 \n",
|
||
"\n",
|
||
" url bedrooms \\\n",
|
||
"0 https://www.rightmove.co.uk/properties/102360773 1 \n",
|
||
"1 https://www.rightmove.co.uk/properties/105836849 3 \n",
|
||
"2 https://www.rightmove.co.uk/properties/107233214 3 \n",
|
||
"3 https://www.rightmove.co.uk/properties/107976896 3 \n",
|
||
"4 https://www.rightmove.co.uk/properties/115499441 3 \n",
|
||
"... ... ... \n",
|
||
"2181 https://www.rightmove.co.uk/properties/86813508 3 \n",
|
||
"2182 https://www.rightmove.co.uk/properties/86813523 3 \n",
|
||
"2183 https://www.rightmove.co.uk/properties/86814669 3 \n",
|
||
"2184 https://www.rightmove.co.uk/properties/86955958 3 \n",
|
||
"2185 https://www.rightmove.co.uk/properties/86972726 2 \n",
|
||
"\n",
|
||
" travel_time_fastest \\\n",
|
||
"0 {'duration': 2695, 'distance': 6467, 'duration... \n",
|
||
"1 {'duration': 2565, 'distance': 14070, 'duratio... \n",
|
||
"2 {'duration': 1714, 'distance': 9570, 'duration... \n",
|
||
"3 {'duration': 1862, 'distance': 8278, 'duration... \n",
|
||
"4 {'duration': 2943, 'distance': 7437, 'duration... \n",
|
||
"... ... \n",
|
||
"2181 {'duration': 2400, 'distance': 13983, 'duratio... \n",
|
||
"2182 {'duration': 2400, 'distance': 13983, 'duratio... \n",
|
||
"2183 {'duration': 2310, 'distance': 12972, 'duratio... \n",
|
||
"2184 {'duration': 2332, 'distance': 6898, 'duration... \n",
|
||
"2185 {'duration': 1393, 'distance': 6390, 'duration... \n",
|
||
"\n",
|
||
" travel_time_second lease_left \\\n",
|
||
"0 {'duration': 1682, 'distance': 6810, 'duration... 119.0 \n",
|
||
"1 {'duration': 2565, 'distance': 14070, 'duratio... NaN \n",
|
||
"2 {'duration': 1774, 'distance': 9570, 'duration... 91.0 \n",
|
||
"3 {'duration': 1862, 'distance': 8278, 'duration... NaN \n",
|
||
"4 {'duration': 2167, 'distance': 9920, 'duration... NaN \n",
|
||
"... ... ... \n",
|
||
"2181 {'duration': 2605, 'distance': 14702, 'duratio... NaN \n",
|
||
"2182 {'duration': 2605, 'distance': 14702, 'duratio... NaN \n",
|
||
"2183 {'duration': 2322, 'distance': 12104, 'duratio... 0.0 \n",
|
||
"2184 {'duration': 2248, 'distance': 6893, 'duration... 105.0 \n",
|
||
"2185 {'duration': 2733, 'distance': 6490, 'duration... 993.0 \n",
|
||
"\n",
|
||
" development decision \n",
|
||
"0 False None \n",
|
||
"1 False None \n",
|
||
"2 False None \n",
|
||
"3 False None \n",
|
||
"4 False None \n",
|
||
"... ... ... \n",
|
||
"2181 True None \n",
|
||
"2182 True None \n",
|
||
"2183 False None \n",
|
||
"2184 False None \n",
|
||
"2185 False None \n",
|
||
"\n",
|
||
"[2186 rows x 11 columns]"
|
||
]
|
||
},
|
||
"execution_count": 4,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"df = pd.DataFrame(ds)\n",
|
||
"df.loc[:, 'decision'] = df.identifier.apply(lambda x: decisions.get(x))\n",
|
||
"df"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 5,
|
||
"id": "d80d9911-9a6d-4608-a6da-11dc864ee32b",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/plain": [
|
||
"(2186, 11)"
|
||
]
|
||
},
|
||
"execution_count": 5,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"df.shape"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 6,
|
||
"id": "7b37ad6b-9b0a-444e-b8c3-6fe4e43e42cb",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/html": [
|
||
"<div>\n",
|
||
"<style scoped>\n",
|
||
" .dataframe tbody tr th:only-of-type {\n",
|
||
" vertical-align: middle;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe tbody tr th {\n",
|
||
" vertical-align: top;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe thead th {\n",
|
||
" text-align: right;\n",
|
||
" }\n",
|
||
"</style>\n",
|
||
"<table border=\"1\" class=\"dataframe\">\n",
|
||
" <thead>\n",
|
||
" <tr style=\"text-align: right;\">\n",
|
||
" <th></th>\n",
|
||
" <th>identifier</th>\n",
|
||
" <th>sqm_ocr</th>\n",
|
||
" <th>price</th>\n",
|
||
" <th>price_per_sqm</th>\n",
|
||
" <th>url</th>\n",
|
||
" <th>bedrooms</th>\n",
|
||
" <th>lease_left</th>\n",
|
||
" <th>development</th>\n",
|
||
" <th>decision</th>\n",
|
||
" <th>a_duration</th>\n",
|
||
" <th>a_initial_walk_duration</th>\n",
|
||
" <th>a_duration_per_transit</th>\n",
|
||
" <th>a_number_of_transit_stops</th>\n",
|
||
" <th>b_duration</th>\n",
|
||
" <th>b_initial_walk_duration</th>\n",
|
||
" <th>b_duration_per_transit</th>\n",
|
||
" <th>b_number_of_transit_stops</th>\n",
|
||
" <th>min_duration</th>\n",
|
||
" </tr>\n",
|
||
" </thead>\n",
|
||
" <tbody>\n",
|
||
" <tr>\n",
|
||
" <th>0</th>\n",
|
||
" <td>102360773</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>350000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/102360773</td>\n",
|
||
" <td>1</td>\n",
|
||
" <td>119.0</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" <td>2695</td>\n",
|
||
" <td>174</td>\n",
|
||
" <td>{'WALK': 414, 'TRANSIT': 2280}</td>\n",
|
||
" <td>1</td>\n",
|
||
" <td>1682.0</td>\n",
|
||
" <td>204.0</td>\n",
|
||
" <td>{'WALK': 608, 'TRANSIT': 804}</td>\n",
|
||
" <td>2.0</td>\n",
|
||
" <td>28.0</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>1</th>\n",
|
||
" <td>105836849</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>400000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/105836849</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" <td>2565</td>\n",
|
||
" <td>340</td>\n",
|
||
" <td>{'WALK': 750, 'TRANSIT': 1800}</td>\n",
|
||
" <td>1</td>\n",
|
||
" <td>2565.0</td>\n",
|
||
" <td>340.0</td>\n",
|
||
" <td>{'WALK': 750, 'TRANSIT': 1800}</td>\n",
|
||
" <td>1.0</td>\n",
|
||
" <td>43.0</td>\n",
|
||
" </tr>\n",
|
||
" </tbody>\n",
|
||
"</table>\n",
|
||
"</div>"
|
||
],
|
||
"text/plain": [
|
||
" identifier sqm_ocr price price_per_sqm \\\n",
|
||
"0 102360773 NaN 350000.0 NaN \n",
|
||
"1 105836849 NaN 400000.0 NaN \n",
|
||
"\n",
|
||
" url bedrooms lease_left \\\n",
|
||
"0 https://www.rightmove.co.uk/properties/102360773 1 119.0 \n",
|
||
"1 https://www.rightmove.co.uk/properties/105836849 3 NaN \n",
|
||
"\n",
|
||
" development decision a_duration a_initial_walk_duration \\\n",
|
||
"0 False None 2695 174 \n",
|
||
"1 False None 2565 340 \n",
|
||
"\n",
|
||
" a_duration_per_transit a_number_of_transit_stops b_duration \\\n",
|
||
"0 {'WALK': 414, 'TRANSIT': 2280} 1 1682.0 \n",
|
||
"1 {'WALK': 750, 'TRANSIT': 1800} 1 2565.0 \n",
|
||
"\n",
|
||
" b_initial_walk_duration b_duration_per_transit \\\n",
|
||
"0 204.0 {'WALK': 608, 'TRANSIT': 804} \n",
|
||
"1 340.0 {'WALK': 750, 'TRANSIT': 1800} \n",
|
||
"\n",
|
||
" b_number_of_transit_stops min_duration \n",
|
||
"0 2.0 28.0 \n",
|
||
"1 1.0 43.0 "
|
||
]
|
||
},
|
||
"execution_count": 6,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"dropcolumns = ['distance_per_transit', 'duration_static', 'distance']\n",
|
||
"s1 = df['travel_time_fastest'].apply(pd.Series).drop(dropcolumns, axis=1)\n",
|
||
"s1.columns = ['a_' + c for c in s1.columns]\n",
|
||
"\n",
|
||
"s2 = df['travel_time_second'].apply(pd.Series).drop(dropcolumns, axis=1)\n",
|
||
"s2.columns = ['b_' + c for c in s2.columns]\n",
|
||
"\n",
|
||
"df2 = pd.concat([df.drop(['travel_time_fastest', 'travel_time_second'], axis=1), s1, s2], axis=1)\n",
|
||
"df2.loc[:, 'min_duration'] = (df2.loc[:, ['a_duration', 'b_duration']].min(axis=1) / 60).round()\n",
|
||
"df2.head(2)"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 7,
|
||
"id": "8c75aaa6-6113-482f-809b-11e405510184",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"df2.to_clipboard()"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 8,
|
||
"id": "79f99692-91e8-4915-9b57-7b3a1efd7d3a",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/html": [
|
||
"<div>\n",
|
||
"<style scoped>\n",
|
||
" .dataframe tbody tr th:only-of-type {\n",
|
||
" vertical-align: middle;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe tbody tr th {\n",
|
||
" vertical-align: top;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe thead th {\n",
|
||
" text-align: right;\n",
|
||
" }\n",
|
||
"</style>\n",
|
||
"<table border=\"1\" class=\"dataframe\">\n",
|
||
" <thead>\n",
|
||
" <tr style=\"text-align: right;\">\n",
|
||
" <th></th>\n",
|
||
" <th>identifier</th>\n",
|
||
" <th>sqm_ocr</th>\n",
|
||
" <th>price</th>\n",
|
||
" <th>price_per_sqm</th>\n",
|
||
" <th>url</th>\n",
|
||
" <th>bedrooms</th>\n",
|
||
" <th>lease_left</th>\n",
|
||
" <th>development</th>\n",
|
||
" <th>decision</th>\n",
|
||
" <th>duration</th>\n",
|
||
" <th>initial_walk_duration</th>\n",
|
||
" <th>duration_per_transit</th>\n",
|
||
" <th>number_of_transit_stops</th>\n",
|
||
" </tr>\n",
|
||
" </thead>\n",
|
||
" <tbody>\n",
|
||
" <tr>\n",
|
||
" <th>0</th>\n",
|
||
" <td>102360773</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>350000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/102360773</td>\n",
|
||
" <td>1</td>\n",
|
||
" <td>119.0</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" <td>45</td>\n",
|
||
" <td>174</td>\n",
|
||
" <td>{'WALK': 414, 'TRANSIT': 2280}</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>1</th>\n",
|
||
" <td>105836849</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>400000.0</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/105836849</td>\n",
|
||
" <td>3</td>\n",
|
||
" <td>NaN</td>\n",
|
||
" <td>False</td>\n",
|
||
" <td>None</td>\n",
|
||
" <td>43</td>\n",
|
||
" <td>340</td>\n",
|
||
" <td>{'WALK': 750, 'TRANSIT': 1800}</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" </tbody>\n",
|
||
"</table>\n",
|
||
"</div>"
|
||
],
|
||
"text/plain": [
|
||
" identifier sqm_ocr price price_per_sqm \\\n",
|
||
"0 102360773 NaN 350000.0 NaN \n",
|
||
"1 105836849 NaN 400000.0 NaN \n",
|
||
"\n",
|
||
" url bedrooms lease_left \\\n",
|
||
"0 https://www.rightmove.co.uk/properties/102360773 1 119.0 \n",
|
||
"1 https://www.rightmove.co.uk/properties/105836849 3 NaN \n",
|
||
"\n",
|
||
" development decision duration initial_walk_duration \\\n",
|
||
"0 False None 45 174 \n",
|
||
"1 False None 43 340 \n",
|
||
"\n",
|
||
" duration_per_transit number_of_transit_stops \n",
|
||
"0 {'WALK': 414, 'TRANSIT': 2280} 1 \n",
|
||
"1 {'WALK': 750, 'TRANSIT': 1800} 1 "
|
||
]
|
||
},
|
||
"execution_count": 8,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"dropcolumns = ['distance_per_transit', 'duration_static', 'distance']\n",
|
||
"s1 = df['travel_time_fastest'].apply(pd.Series).drop(dropcolumns, axis=1)\n",
|
||
"\n",
|
||
"df3 = pd.concat([df.drop(['travel_time_fastest', 'travel_time_second'], axis=1), s1], axis=1)\n",
|
||
"df3.loc[:, 'duration'] = (df3.loc[:, ['duration']].min(axis=1) / 60).round()\n",
|
||
"df3.to_clipboard()\n",
|
||
"df3.head(2)"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "abcbde40-7432-4449-957a-79ce2ca126eb",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 5,
|
||
"id": "99c5b304-3d13-466b-a9f5-83a5db6311b5",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/html": [
|
||
"<div>\n",
|
||
"<style scoped>\n",
|
||
" .dataframe tbody tr th:only-of-type {\n",
|
||
" vertical-align: middle;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe tbody tr th {\n",
|
||
" vertical-align: top;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe thead th {\n",
|
||
" text-align: right;\n",
|
||
" }\n",
|
||
"</style>\n",
|
||
"<table border=\"1\" class=\"dataframe\">\n",
|
||
" <thead>\n",
|
||
" <tr style=\"text-align: right;\">\n",
|
||
" <th></th>\n",
|
||
" <th>sqm_ocr</th>\n",
|
||
" <th>price</th>\n",
|
||
" <th>price_per_sqm</th>\n",
|
||
" <th>url</th>\n",
|
||
" <th>bedrooms</th>\n",
|
||
" </tr>\n",
|
||
" </thead>\n",
|
||
" <tbody>\n",
|
||
" <tr>\n",
|
||
" <th>953</th>\n",
|
||
" <td>74142.40</td>\n",
|
||
" <td>550000.0</td>\n",
|
||
" <td>7.418157</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145546538</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>249</th>\n",
|
||
" <td>21850.10</td>\n",
|
||
" <td>725000.0</td>\n",
|
||
" <td>33.180626</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/140821736</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>177</th>\n",
|
||
" <td>137.50</td>\n",
|
||
" <td>695000.0</td>\n",
|
||
" <td>5054.545455</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/139354259</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>352</th>\n",
|
||
" <td>85.50</td>\n",
|
||
" <td>475000.0</td>\n",
|
||
" <td>5555.555556</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/142142348</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>36</th>\n",
|
||
" <td>82.90</td>\n",
|
||
" <td>475000.0</td>\n",
|
||
" <td>5729.794934</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/128925950</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>700</th>\n",
|
||
" <td>101.00</td>\n",
|
||
" <td>650000.0</td>\n",
|
||
" <td>6435.643564</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144591572</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>823</th>\n",
|
||
" <td>80.40</td>\n",
|
||
" <td>525000.0</td>\n",
|
||
" <td>6529.850746</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145051769</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>214</th>\n",
|
||
" <td>91.51</td>\n",
|
||
" <td>650000.0</td>\n",
|
||
" <td>7103.048847</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/140326307</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>171</th>\n",
|
||
" <td>91.10</td>\n",
|
||
" <td>650000.0</td>\n",
|
||
" <td>7135.016465</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/139245428</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>598</th>\n",
|
||
" <td>109.40</td>\n",
|
||
" <td>795000.0</td>\n",
|
||
" <td>7266.910420</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144034655</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>146</th>\n",
|
||
" <td>93.78</td>\n",
|
||
" <td>700000.0</td>\n",
|
||
" <td>7464.278098</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/138510395</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>711</th>\n",
|
||
" <td>100.31</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>7476.821852</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144620303</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>592</th>\n",
|
||
" <td>86.49</td>\n",
|
||
" <td>650000.0</td>\n",
|
||
" <td>7515.319690</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143987669</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>624</th>\n",
|
||
" <td>101.35</td>\n",
|
||
" <td>775000.0</td>\n",
|
||
" <td>7646.768624</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144217922</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>55</th>\n",
|
||
" <td>97.60</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>7684.426230</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/132564737</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>851</th>\n",
|
||
" <td>81.01</td>\n",
|
||
" <td>625000.0</td>\n",
|
||
" <td>7715.096902</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145172504</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>356</th>\n",
|
||
" <td>89.00</td>\n",
|
||
" <td>695000.0</td>\n",
|
||
" <td>7808.988764</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/142185623</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>748</th>\n",
|
||
" <td>85.90</td>\n",
|
||
" <td>675000.0</td>\n",
|
||
" <td>7857.974389</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144771281</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>204</th>\n",
|
||
" <td>82.40</td>\n",
|
||
" <td>650000.0</td>\n",
|
||
" <td>7888.349515</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/140173319</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>654</th>\n",
|
||
" <td>83.33</td>\n",
|
||
" <td>665000.0</td>\n",
|
||
" <td>7980.319213</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144361100</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>514</th>\n",
|
||
" <td>87.50</td>\n",
|
||
" <td>700000.0</td>\n",
|
||
" <td>8000.000000</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143460365</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>762</th>\n",
|
||
" <td>92.90</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>8073.196986</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144846725</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>963</th>\n",
|
||
" <td>92.90</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>8073.196986</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145565252</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>938</th>\n",
|
||
" <td>86.10</td>\n",
|
||
" <td>699000.0</td>\n",
|
||
" <td>8118.466899</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145491137</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>403</th>\n",
|
||
" <td>90.60</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>8278.145695</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/142704416</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>302</th>\n",
|
||
" <td>93.30</td>\n",
|
||
" <td>795000.0</td>\n",
|
||
" <td>8520.900322</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141595433</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>671</th>\n",
|
||
" <td>93.27</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>8577.248847</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144429140</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>349</th>\n",
|
||
" <td>80.57</td>\n",
|
||
" <td>695000.0</td>\n",
|
||
" <td>8626.039469</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/142115918</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>224</th>\n",
|
||
" <td>90.10</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>8879.023307</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/140464481</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>740</th>\n",
|
||
" <td>81.70</td>\n",
|
||
" <td>735000.0</td>\n",
|
||
" <td>8996.328029</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144722414</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>326</th>\n",
|
||
" <td>87.30</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9163.802978</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141846023</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>319</th>\n",
|
||
" <td>81.57</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>9194.556822</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141797357</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>558</th>\n",
|
||
" <td>86.77</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9219.776420</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143758763</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>712</th>\n",
|
||
" <td>86.30</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9269.988413</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144622157</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>473</th>\n",
|
||
" <td>86.00</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9302.325581</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143210102</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>105</th>\n",
|
||
" <td>80.40</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>9328.358209</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/136988726</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>1003</th>\n",
|
||
" <td>80.27</td>\n",
|
||
" <td>750000.0</td>\n",
|
||
" <td>9343.465803</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86775291</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>235</th>\n",
|
||
" <td>82.80</td>\n",
|
||
" <td>775000.0</td>\n",
|
||
" <td>9359.903382</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/140611055</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>65</th>\n",
|
||
" <td>85.10</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9400.705053</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/134116232</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>30</th>\n",
|
||
" <td>83.70</td>\n",
|
||
" <td>795000.0</td>\n",
|
||
" <td>9498.207885</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/127787960</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>1025</th>\n",
|
||
" <td>81.60</td>\n",
|
||
" <td>790000.0</td>\n",
|
||
" <td>9681.372549</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86972726</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>88</th>\n",
|
||
" <td>81.75</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9785.932722</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/136012193</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>454</th>\n",
|
||
" <td>80.80</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9900.990099</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143138867</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>343</th>\n",
|
||
" <td>80.64</td>\n",
|
||
" <td>800000.0</td>\n",
|
||
" <td>9920.634921</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/142032935</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" </tbody>\n",
|
||
"</table>\n",
|
||
"</div>"
|
||
],
|
||
"text/plain": [
|
||
" sqm_ocr price price_per_sqm \\\n",
|
||
"953 74142.40 550000.0 7.418157 \n",
|
||
"249 21850.10 725000.0 33.180626 \n",
|
||
"177 137.50 695000.0 5054.545455 \n",
|
||
"352 85.50 475000.0 5555.555556 \n",
|
||
"36 82.90 475000.0 5729.794934 \n",
|
||
"700 101.00 650000.0 6435.643564 \n",
|
||
"823 80.40 525000.0 6529.850746 \n",
|
||
"214 91.51 650000.0 7103.048847 \n",
|
||
"171 91.10 650000.0 7135.016465 \n",
|
||
"598 109.40 795000.0 7266.910420 \n",
|
||
"146 93.78 700000.0 7464.278098 \n",
|
||
"711 100.31 750000.0 7476.821852 \n",
|
||
"592 86.49 650000.0 7515.319690 \n",
|
||
"624 101.35 775000.0 7646.768624 \n",
|
||
"55 97.60 750000.0 7684.426230 \n",
|
||
"851 81.01 625000.0 7715.096902 \n",
|
||
"356 89.00 695000.0 7808.988764 \n",
|
||
"748 85.90 675000.0 7857.974389 \n",
|
||
"204 82.40 650000.0 7888.349515 \n",
|
||
"654 83.33 665000.0 7980.319213 \n",
|
||
"514 87.50 700000.0 8000.000000 \n",
|
||
"762 92.90 750000.0 8073.196986 \n",
|
||
"963 92.90 750000.0 8073.196986 \n",
|
||
"938 86.10 699000.0 8118.466899 \n",
|
||
"403 90.60 750000.0 8278.145695 \n",
|
||
"302 93.30 795000.0 8520.900322 \n",
|
||
"671 93.27 800000.0 8577.248847 \n",
|
||
"349 80.57 695000.0 8626.039469 \n",
|
||
"224 90.10 800000.0 8879.023307 \n",
|
||
"740 81.70 735000.0 8996.328029 \n",
|
||
"326 87.30 800000.0 9163.802978 \n",
|
||
"319 81.57 750000.0 9194.556822 \n",
|
||
"558 86.77 800000.0 9219.776420 \n",
|
||
"712 86.30 800000.0 9269.988413 \n",
|
||
"473 86.00 800000.0 9302.325581 \n",
|
||
"105 80.40 750000.0 9328.358209 \n",
|
||
"1003 80.27 750000.0 9343.465803 \n",
|
||
"235 82.80 775000.0 9359.903382 \n",
|
||
"65 85.10 800000.0 9400.705053 \n",
|
||
"30 83.70 795000.0 9498.207885 \n",
|
||
"1025 81.60 790000.0 9681.372549 \n",
|
||
"88 81.75 800000.0 9785.932722 \n",
|
||
"454 80.80 800000.0 9900.990099 \n",
|
||
"343 80.64 800000.0 9920.634921 \n",
|
||
"\n",
|
||
" url bedrooms \n",
|
||
"953 https://www.rightmove.co.uk/properties/145546538 2 \n",
|
||
"249 https://www.rightmove.co.uk/properties/140821736 2 \n",
|
||
"177 https://www.rightmove.co.uk/properties/139354259 2 \n",
|
||
"352 https://www.rightmove.co.uk/properties/142142348 2 \n",
|
||
"36 https://www.rightmove.co.uk/properties/128925950 2 \n",
|
||
"700 https://www.rightmove.co.uk/properties/144591572 1 \n",
|
||
"823 https://www.rightmove.co.uk/properties/145051769 2 \n",
|
||
"214 https://www.rightmove.co.uk/properties/140326307 2 \n",
|
||
"171 https://www.rightmove.co.uk/properties/139245428 2 \n",
|
||
"598 https://www.rightmove.co.uk/properties/144034655 2 \n",
|
||
"146 https://www.rightmove.co.uk/properties/138510395 2 \n",
|
||
"711 https://www.rightmove.co.uk/properties/144620303 2 \n",
|
||
"592 https://www.rightmove.co.uk/properties/143987669 2 \n",
|
||
"624 https://www.rightmove.co.uk/properties/144217922 1 \n",
|
||
"55 https://www.rightmove.co.uk/properties/132564737 2 \n",
|
||
"851 https://www.rightmove.co.uk/properties/145172504 2 \n",
|
||
"356 https://www.rightmove.co.uk/properties/142185623 2 \n",
|
||
"748 https://www.rightmove.co.uk/properties/144771281 2 \n",
|
||
"204 https://www.rightmove.co.uk/properties/140173319 2 \n",
|
||
"654 https://www.rightmove.co.uk/properties/144361100 2 \n",
|
||
"514 https://www.rightmove.co.uk/properties/143460365 1 \n",
|
||
"762 https://www.rightmove.co.uk/properties/144846725 1 \n",
|
||
"963 https://www.rightmove.co.uk/properties/145565252 2 \n",
|
||
"938 https://www.rightmove.co.uk/properties/145491137 2 \n",
|
||
"403 https://www.rightmove.co.uk/properties/142704416 2 \n",
|
||
"302 https://www.rightmove.co.uk/properties/141595433 2 \n",
|
||
"671 https://www.rightmove.co.uk/properties/144429140 2 \n",
|
||
"349 https://www.rightmove.co.uk/properties/142115918 2 \n",
|
||
"224 https://www.rightmove.co.uk/properties/140464481 2 \n",
|
||
"740 https://www.rightmove.co.uk/properties/144722414 2 \n",
|
||
"326 https://www.rightmove.co.uk/properties/141846023 2 \n",
|
||
"319 https://www.rightmove.co.uk/properties/141797357 2 \n",
|
||
"558 https://www.rightmove.co.uk/properties/143758763 2 \n",
|
||
"712 https://www.rightmove.co.uk/properties/144622157 2 \n",
|
||
"473 https://www.rightmove.co.uk/properties/143210102 2 \n",
|
||
"105 https://www.rightmove.co.uk/properties/136988726 2 \n",
|
||
"1003 https://www.rightmove.co.uk/properties/86775291 2 \n",
|
||
"235 https://www.rightmove.co.uk/properties/140611055 2 \n",
|
||
"65 https://www.rightmove.co.uk/properties/134116232 2 \n",
|
||
"30 https://www.rightmove.co.uk/properties/127787960 1 \n",
|
||
"1025 https://www.rightmove.co.uk/properties/86972726 2 \n",
|
||
"88 https://www.rightmove.co.uk/properties/136012193 1 \n",
|
||
"454 https://www.rightmove.co.uk/properties/143138867 2 \n",
|
||
"343 https://www.rightmove.co.uk/properties/142032935 2 "
|
||
]
|
||
},
|
||
"execution_count": 5,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"df2 = df[df.sqm_ocr > 80]\n",
|
||
"df2.sort_values('price_per_sqm')"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 6,
|
||
"id": "52545cfa-0932-46fe-ba7e-961fd43f2786",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/html": [
|
||
"<div>\n",
|
||
"<style scoped>\n",
|
||
" .dataframe tbody tr th:only-of-type {\n",
|
||
" vertical-align: middle;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe tbody tr th {\n",
|
||
" vertical-align: top;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe thead th {\n",
|
||
" text-align: right;\n",
|
||
" }\n",
|
||
"</style>\n",
|
||
"<table border=\"1\" class=\"dataframe\">\n",
|
||
" <thead>\n",
|
||
" <tr style=\"text-align: right;\">\n",
|
||
" <th></th>\n",
|
||
" <th>sqm_ocr</th>\n",
|
||
" <th>price</th>\n",
|
||
" <th>price_per_sqm</th>\n",
|
||
" <th>url</th>\n",
|
||
" <th>bedrooms</th>\n",
|
||
" </tr>\n",
|
||
" </thead>\n",
|
||
" <tbody>\n",
|
||
" <tr>\n",
|
||
" <th>359</th>\n",
|
||
" <td>73.40</td>\n",
|
||
" <td>400000.0</td>\n",
|
||
" <td>5449.591281</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/142186991</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>293</th>\n",
|
||
" <td>77.00</td>\n",
|
||
" <td>425000.0</td>\n",
|
||
" <td>5519.480519</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141437783</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>352</th>\n",
|
||
" <td>85.50</td>\n",
|
||
" <td>475000.0</td>\n",
|
||
" <td>5555.555556</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/142142348</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>685</th>\n",
|
||
" <td>76.03</td>\n",
|
||
" <td>425000.0</td>\n",
|
||
" <td>5589.898724</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144494012</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>36</th>\n",
|
||
" <td>82.90</td>\n",
|
||
" <td>475000.0</td>\n",
|
||
" <td>5729.794934</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/128925950</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>491</th>\n",
|
||
" <td>71.68</td>\n",
|
||
" <td>450000.0</td>\n",
|
||
" <td>6277.901786</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143315840</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>1020</th>\n",
|
||
" <td>73.67</td>\n",
|
||
" <td>495000.0</td>\n",
|
||
" <td>6719.152980</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86807916</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" </tbody>\n",
|
||
"</table>\n",
|
||
"</div>"
|
||
],
|
||
"text/plain": [
|
||
" sqm_ocr price price_per_sqm \\\n",
|
||
"359 73.40 400000.0 5449.591281 \n",
|
||
"293 77.00 425000.0 5519.480519 \n",
|
||
"352 85.50 475000.0 5555.555556 \n",
|
||
"685 76.03 425000.0 5589.898724 \n",
|
||
"36 82.90 475000.0 5729.794934 \n",
|
||
"491 71.68 450000.0 6277.901786 \n",
|
||
"1020 73.67 495000.0 6719.152980 \n",
|
||
"\n",
|
||
" url bedrooms \n",
|
||
"359 https://www.rightmove.co.uk/properties/142186991 2 \n",
|
||
"293 https://www.rightmove.co.uk/properties/141437783 2 \n",
|
||
"352 https://www.rightmove.co.uk/properties/142142348 2 \n",
|
||
"685 https://www.rightmove.co.uk/properties/144494012 2 \n",
|
||
"36 https://www.rightmove.co.uk/properties/128925950 2 \n",
|
||
"491 https://www.rightmove.co.uk/properties/143315840 2 \n",
|
||
"1020 https://www.rightmove.co.uk/properties/86807916 2 "
|
||
]
|
||
},
|
||
"execution_count": 6,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"df2 = df[(df.sqm_ocr > 70) & (df.price <500000)]\n",
|
||
"df2.sort_values('price_per_sqm')"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 7,
|
||
"id": "d0246926-13ef-4110-8e3a-fb676a55c2a6",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/html": [
|
||
"<div>\n",
|
||
"<style scoped>\n",
|
||
" .dataframe tbody tr th:only-of-type {\n",
|
||
" vertical-align: middle;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe tbody tr th {\n",
|
||
" vertical-align: top;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe thead th {\n",
|
||
" text-align: right;\n",
|
||
" }\n",
|
||
"</style>\n",
|
||
"<table border=\"1\" class=\"dataframe\">\n",
|
||
" <thead>\n",
|
||
" <tr style=\"text-align: right;\">\n",
|
||
" <th></th>\n",
|
||
" <th>sqm_ocr</th>\n",
|
||
" <th>price</th>\n",
|
||
" <th>price_per_sqm</th>\n",
|
||
" <th>url</th>\n",
|
||
" <th>bedrooms</th>\n",
|
||
" </tr>\n",
|
||
" </thead>\n",
|
||
" <tbody>\n",
|
||
" <tr>\n",
|
||
" <th>953</th>\n",
|
||
" <td>74142.40</td>\n",
|
||
" <td>550000.0</td>\n",
|
||
" <td>7.418157</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145546538</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>823</th>\n",
|
||
" <td>80.40</td>\n",
|
||
" <td>525000.0</td>\n",
|
||
" <td>6529.850746</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145051769</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>492</th>\n",
|
||
" <td>73.90</td>\n",
|
||
" <td>525000.0</td>\n",
|
||
" <td>7104.194858</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143317361</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>561</th>\n",
|
||
" <td>76.70</td>\n",
|
||
" <td>550000.0</td>\n",
|
||
" <td>7170.795306</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143780789</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>69</th>\n",
|
||
" <td>73.10</td>\n",
|
||
" <td>525000.0</td>\n",
|
||
" <td>7181.942544</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/134574563</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>988</th>\n",
|
||
" <td>72.40</td>\n",
|
||
" <td>525000.0</td>\n",
|
||
" <td>7251.381215</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86648925</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>272</th>\n",
|
||
" <td>77.89</td>\n",
|
||
" <td>575000.0</td>\n",
|
||
" <td>7382.205675</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141131297</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>521</th>\n",
|
||
" <td>70.88</td>\n",
|
||
" <td>525000.0</td>\n",
|
||
" <td>7406.884876</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143514149</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>324</th>\n",
|
||
" <td>76.10</td>\n",
|
||
" <td>575000.0</td>\n",
|
||
" <td>7555.847569</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141831353</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>1018</th>\n",
|
||
" <td>71.80</td>\n",
|
||
" <td>550000.0</td>\n",
|
||
" <td>7660.167131</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86804832</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>526</th>\n",
|
||
" <td>78.00</td>\n",
|
||
" <td>600000.0</td>\n",
|
||
" <td>7692.307692</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/143552156</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>817</th>\n",
|
||
" <td>71.17</td>\n",
|
||
" <td>550000.0</td>\n",
|
||
" <td>7727.975270</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145035929</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>843</th>\n",
|
||
" <td>77.50</td>\n",
|
||
" <td>600000.0</td>\n",
|
||
" <td>7741.935484</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145144988</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>337</th>\n",
|
||
" <td>70.60</td>\n",
|
||
" <td>550000.0</td>\n",
|
||
" <td>7790.368272</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141904286</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>233</th>\n",
|
||
" <td>75.70</td>\n",
|
||
" <td>600000.0</td>\n",
|
||
" <td>7926.023778</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/140582213</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>763</th>\n",
|
||
" <td>75.00</td>\n",
|
||
" <td>600000.0</td>\n",
|
||
" <td>8000.000000</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/144862070</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>315</th>\n",
|
||
" <td>71.30</td>\n",
|
||
" <td>590000.0</td>\n",
|
||
" <td>8274.894811</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/141669686</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>899</th>\n",
|
||
" <td>71.47</td>\n",
|
||
" <td>595000.0</td>\n",
|
||
" <td>8325.171401</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/145362911</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>147</th>\n",
|
||
" <td>71.50</td>\n",
|
||
" <td>600000.0</td>\n",
|
||
" <td>8391.608392</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/138537527</td>\n",
|
||
" <td>1</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>973</th>\n",
|
||
" <td>70.89</td>\n",
|
||
" <td>595000.0</td>\n",
|
||
" <td>8393.285372</td>\n",
|
||
" <td>https://www.rightmove.co.uk/properties/86296491</td>\n",
|
||
" <td>2</td>\n",
|
||
" </tr>\n",
|
||
" </tbody>\n",
|
||
"</table>\n",
|
||
"</div>"
|
||
],
|
||
"text/plain": [
|
||
" sqm_ocr price price_per_sqm \\\n",
|
||
"953 74142.40 550000.0 7.418157 \n",
|
||
"823 80.40 525000.0 6529.850746 \n",
|
||
"492 73.90 525000.0 7104.194858 \n",
|
||
"561 76.70 550000.0 7170.795306 \n",
|
||
"69 73.10 525000.0 7181.942544 \n",
|
||
"988 72.40 525000.0 7251.381215 \n",
|
||
"272 77.89 575000.0 7382.205675 \n",
|
||
"521 70.88 525000.0 7406.884876 \n",
|
||
"324 76.10 575000.0 7555.847569 \n",
|
||
"1018 71.80 550000.0 7660.167131 \n",
|
||
"526 78.00 600000.0 7692.307692 \n",
|
||
"817 71.17 550000.0 7727.975270 \n",
|
||
"843 77.50 600000.0 7741.935484 \n",
|
||
"337 70.60 550000.0 7790.368272 \n",
|
||
"233 75.70 600000.0 7926.023778 \n",
|
||
"763 75.00 600000.0 8000.000000 \n",
|
||
"315 71.30 590000.0 8274.894811 \n",
|
||
"899 71.47 595000.0 8325.171401 \n",
|
||
"147 71.50 600000.0 8391.608392 \n",
|
||
"973 70.89 595000.0 8393.285372 \n",
|
||
"\n",
|
||
" url bedrooms \n",
|
||
"953 https://www.rightmove.co.uk/properties/145546538 2 \n",
|
||
"823 https://www.rightmove.co.uk/properties/145051769 2 \n",
|
||
"492 https://www.rightmove.co.uk/properties/143317361 2 \n",
|
||
"561 https://www.rightmove.co.uk/properties/143780789 2 \n",
|
||
"69 https://www.rightmove.co.uk/properties/134574563 2 \n",
|
||
"988 https://www.rightmove.co.uk/properties/86648925 2 \n",
|
||
"272 https://www.rightmove.co.uk/properties/141131297 2 \n",
|
||
"521 https://www.rightmove.co.uk/properties/143514149 2 \n",
|
||
"324 https://www.rightmove.co.uk/properties/141831353 2 \n",
|
||
"1018 https://www.rightmove.co.uk/properties/86804832 2 \n",
|
||
"526 https://www.rightmove.co.uk/properties/143552156 1 \n",
|
||
"817 https://www.rightmove.co.uk/properties/145035929 2 \n",
|
||
"843 https://www.rightmove.co.uk/properties/145144988 2 \n",
|
||
"337 https://www.rightmove.co.uk/properties/141904286 2 \n",
|
||
"233 https://www.rightmove.co.uk/properties/140582213 2 \n",
|
||
"763 https://www.rightmove.co.uk/properties/144862070 2 \n",
|
||
"315 https://www.rightmove.co.uk/properties/141669686 2 \n",
|
||
"899 https://www.rightmove.co.uk/properties/145362911 2 \n",
|
||
"147 https://www.rightmove.co.uk/properties/138537527 1 \n",
|
||
"973 https://www.rightmove.co.uk/properties/86296491 2 "
|
||
]
|
||
},
|
||
"execution_count": 7,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"df2 = df[(df.sqm_ocr > 70) & (df.price <=600000)& (df.price >500000)]\n",
|
||
"df2.sort_values('price_per_sqm')"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "b9f2ca7b-8476-47f1-a34f-7acdfbe84bd1",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": []
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 15,
|
||
"id": "edd9fa24-cad2-4448-9b17-c6d514564f41",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"from data_access import Listing\n",
|
||
"import pytesseract\n",
|
||
"from PIL import Image\n",
|
||
"import re"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 18,
|
||
"id": "002b2a3a-3ecc-45c1-8c2f-c143380ee0d5",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"[('3.14', 'm'), ('3.43', 'm'), ('3.89', 'm'), ('3.62', 'm'), ('3.89', 'm'), ('2.88', 'm'), ('75.3', 'sq. m'), ('4.07', 'm'), ('6.18', 'm'), ('2.79', 'm'), ('3.34', 'm'), ('2.79', 'm'), ('4.34', 'm'), ('1.76', 'm'), ('2.29', 'm'), ('1.76', 'm'), ('1.92', 'm'), ('75.3', 'sq. m')]\n",
|
||
"-----\n",
|
||
"Bedroom 2\n",
|
||
"Bedroom 1 3.14m x 3.43m\n",
|
||
"3.89m x 3.62m (10'4\" x 11'3\")\n",
|
||
"(12'9\" x 11'10\")\n",
|
||
"\n",
|
||
"Bedroom 3\n",
|
||
"3.89m x 2.88m\n",
|
||
"(12'9\" x 9'5\")\n",
|
||
"\n",
|
||
"xt\n",
|
||
"\n",
|
||
"v\n",
|
||
"\n",
|
||
"Levita House NW1\n",
|
||
"\n",
|
||
"Approx. 75.3 sq. metres (810.2 sq. feet)\n",
|
||
"\n",
|
||
"Storage\n",
|
||
"\n",
|
||
". Hall\n",
|
||
"Kitchen 4.07m x 6.18m\n",
|
||
"2.79m x 3.34m (13'4\" x 20'3\") Bedroom 4\n",
|
||
"\n",
|
||
"(9'2\" x 10'11\") 2.79m x 4.34m\n",
|
||
"\n",
|
||
"(9'2\" x 14'3\")\n",
|
||
"\n",
|
||
"Bathroom 1\n",
|
||
"1.76m x 2.29m\n",
|
||
"(5'9\" x 76\")\n",
|
||
"\n",
|
||
"Bathroom 2\n",
|
||
"1.76m x 1.92m\n",
|
||
"(5'9\" x 64\")\n",
|
||
"\n",
|
||
"Total area: approx. 75.3 sq. metres (810.2 sq. feet)\n",
|
||
"\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"l = Listing(144497822)\n",
|
||
"for path in l.path_floorplan_folder().iterdir():\n",
|
||
" img = Image.open(path)\n",
|
||
" text = pytesseract.image_to_string(img)\n",
|
||
" sqmregex = r'(\\d+\\.\\d*) ?(sq ?m|sq. ?m|m)'\n",
|
||
" matches = re.findall(sqmregex, text.lower())\n",
|
||
" print(matches)\n",
|
||
" print(\"-----\")\n",
|
||
" print(text)"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 9,
|
||
"id": "015e870d-0cf0-4d07-a9ae-4e80d128b26c",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"import enum"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 17,
|
||
"id": "577dcd5f-07df-4d61-b837-a90db59e3ed5",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"class PropertyType(enum.StrEnum): \n",
|
||
" BUNGALOW= \"bungalow\"\n",
|
||
" DETACHED= \"detached\"\n",
|
||
" FLAT= \"flat\"\n",
|
||
" LAND= \"land\"\n",
|
||
" PARK_HOME= \"park-home\"\n",
|
||
" SEMI_DETACHED= \"semi-detached\"\n",
|
||
" TERRACED= \"terraced\""
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 19,
|
||
"id": "50cc2eb3-1c3b-49b8-86a3-73dd2d151a61",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/plain": [
|
||
"'bungalow,park-home'"
|
||
]
|
||
},
|
||
"execution_count": 19,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"x = [PropertyType.BUNGALOW,PropertyType.PARK_HOME]\n",
|
||
"','.join(x)"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "87ead853-8a71-4de9-98d1-f4f2673a5592",
|
||
"metadata": {},
|
||
"source": [
|
||
"x"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "862e9e52-53fa-4bf9-8e31-7847481d45be",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": []
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "e284aefd-c280-4d82-935c-969b022b6bbc",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": []
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "04bb61d5-cba7-4739-9568-b00342c1b636",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": []
|
||
},
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "a73fba2d-afeb-4194-8421-eff8e84a14e9",
|
||
"metadata": {},
|
||
"source": [
|
||
"# Typeahead"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 20,
|
||
"id": "61844fe2-408d-4b89-995f-c31110a850f6",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/plain": [
|
||
"{'key': 'CAMD',\n",
|
||
" 'term': 'CAMD',\n",
|
||
" 'typeAheadLocations': [{'displayName': 'Camden, North West London',\n",
|
||
" 'locationIdentifier': 'REGION^85261',\n",
|
||
" 'normalisedSearchTerm': 'CAMDEN NORTH WEST LONDON'},\n",
|
||
" {'displayName': 'Camden (London Borough)',\n",
|
||
" 'locationIdentifier': 'REGION^93941',\n",
|
||
" 'normalisedSearchTerm': 'CAMDEN LONDON BOROUGH'},\n",
|
||
" {'displayName': 'Camden Town, North West London',\n",
|
||
" 'locationIdentifier': 'REGION^85262',\n",
|
||
" 'normalisedSearchTerm': 'CAMDEN TOWN NORTH WEST LONDON'},\n",
|
||
" {'displayName': 'Camden Town Station',\n",
|
||
" 'locationIdentifier': 'STATION^1712',\n",
|
||
" 'normalisedSearchTerm': 'CAMDEN TOWN STATION'},\n",
|
||
" {'displayName': 'Camden Road Station',\n",
|
||
" 'locationIdentifier': 'STATION^1709',\n",
|
||
" 'normalisedSearchTerm': 'CAMDEN ROAD STATION'},\n",
|
||
" {'displayName': 'Camden Town, Gosport, Hampshire',\n",
|
||
" 'locationIdentifier': 'REGION^76577',\n",
|
||
" 'normalisedSearchTerm': 'CAMDEN TOWN GOSPORT HAMPSHIRE'},\n",
|
||
" {'displayName': 'Camderry, Omagh, County Tyrone, Northern Ireland',\n",
|
||
" 'locationIdentifier': 'REGION^73327',\n",
|
||
" 'normalisedSearchTerm': 'CAMDERRY OMAGH COUNTY TYRONE NORTHERN IRELAND'}],\n",
|
||
" 'isComplete': True}"
|
||
]
|
||
},
|
||
"execution_count": 20,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"import requests\n",
|
||
"\n",
|
||
"cookies = {\n",
|
||
" 'permuserid': '240330LSEXEOANG04Q2VA3OZCIQ8TTSQ',\n",
|
||
" 'TS019c0ed0': '012f990cd3494097746bc0b10b8d61bc6237319024e80701e0e8d735bd7d62a792529823c3164c771f223b0cade5ec9ae4b7fc001c',\n",
|
||
" 'beta_optin': 'N:36:-1',\n",
|
||
" 'RM_Register': 'C',\n",
|
||
" 'JSESSIONID': '0BE8E261D81387C9BC530DB1A5F28955',\n",
|
||
" 'svr': '3111',\n",
|
||
" 'permuserid': '240330LSEXEOANG04Q2VA3OZCIQ8TTSQ',\n",
|
||
" 'TS01ec61d1': '012f990cd3161dd68e4ed69b9d64f7d4de2356c651edef96201cc7facd11b28ef9338596a2022bc2ed7a56f09c5dee5aa1711de2d1',\n",
|
||
" 'rmsessionid': '7bc54ce6-da97-42cf-8719-4e3e9c53e276',\n",
|
||
" 'TS01821201': '012f990cd35255a563a541cfe06e4a774c129628165af71838ccdb7a17919672962514804459a0d9d9a90fe7b8feeec66145e30b98',\n",
|
||
" 'TS01826437': '012f990cd3161dd68e4ed69b9d64f7d4de2356c651edef96201cc7facd11b28ef9338596a2022bc2ed7a56f09c5dee5aa1711de2d1',\n",
|
||
" 'TPCmaxPrice': '800000',\n",
|
||
" 'TS01a07bd2': '012f990cd3161dd68e4ed69b9d64f7d4de2356c651edef96201cc7facd11b28ef9338596a2022bc2ed7a56f09c5dee5aa1711de2d1',\n",
|
||
" 'TPCminPrice': '700000',\n",
|
||
"}\n",
|
||
"\n",
|
||
"headers = {\n",
|
||
" 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:123.0) Gecko/20100101 Firefox/123.0',\n",
|
||
" 'Accept': 'application/json, text/javascript',\n",
|
||
" 'Accept-Language': 'en-GB,en;q=0.5',\n",
|
||
" # 'Accept-Encoding': 'gzip, deflate, br',\n",
|
||
" 'Referer': 'https://www.rightmove.co.uk/property-for-sale/find.html?locationIdentifier=REGION%5E87515&maxBedrooms=3&minBedrooms=1&maxPrice=800000&minPrice=750001&propertyTypes=&includeSSTC=false&mustHave=&dontShow=retirement%2CsharedOwnership&furnishTypes=&keywords=',\n",
|
||
" 'X-Correlation-Text-Val': 'source=search',\n",
|
||
" 'Content-Type': 'application/x-www-form-urlencoded',\n",
|
||
" 'traceparent': '00-d3020142d839bf1ef2b172fa596acea8-605693e1c4c5cbb5-00',\n",
|
||
" 'DNT': '1',\n",
|
||
" 'Sec-GPC': '1',\n",
|
||
" 'Connection': 'keep-alive',\n",
|
||
" # 'Cookie': 'permuserid=240330LSEXEOANG04Q2VA3OZCIQ8TTSQ; TS019c0ed0=012f990cd3494097746bc0b10b8d61bc6237319024e80701e0e8d735bd7d62a792529823c3164c771f223b0cade5ec9ae4b7fc001c; beta_optin=N:36:-1; RM_Register=C; JSESSIONID=0BE8E261D81387C9BC530DB1A5F28955; svr=3111; permuserid=240330LSEXEOANG04Q2VA3OZCIQ8TTSQ; TS01ec61d1=012f990cd3161dd68e4ed69b9d64f7d4de2356c651edef96201cc7facd11b28ef9338596a2022bc2ed7a56f09c5dee5aa1711de2d1; rmsessionid=7bc54ce6-da97-42cf-8719-4e3e9c53e276; TS01821201=012f990cd35255a563a541cfe06e4a774c129628165af71838ccdb7a17919672962514804459a0d9d9a90fe7b8feeec66145e30b98; TS01826437=012f990cd3161dd68e4ed69b9d64f7d4de2356c651edef96201cc7facd11b28ef9338596a2022bc2ed7a56f09c5dee5aa1711de2d1; TPCmaxPrice=800000; TS01a07bd2=012f990cd3161dd68e4ed69b9d64f7d4de2356c651edef96201cc7facd11b28ef9338596a2022bc2ed7a56f09c5dee5aa1711de2d1; TPCminPrice=700000',\n",
|
||
" 'Sec-Fetch-Dest': 'empty',\n",
|
||
" 'Sec-Fetch-Mode': 'cors',\n",
|
||
" 'Sec-Fetch-Site': 'same-origin',\n",
|
||
"}\n",
|
||
"\n",
|
||
"response = requests.get('https://www.rightmove.co.uk/typeAhead/uknostreet/CA/MD/EN/', cookies=cookies, headers=headers)\n",
|
||
"response.json()"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 39,
|
||
"id": "eba907d0-680d-4374-a56e-95bd4abf0053",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"import requests\n",
|
||
"from textwrap import wrap\n",
|
||
"\n",
|
||
"def query_loc_identifier(name: str)->str:\n",
|
||
" \"\"\"\n",
|
||
" \n",
|
||
" \"\"\"\n",
|
||
" name = name.upper()\n",
|
||
" name = '/'.join(wrap(name,2))\n",
|
||
" \n",
|
||
" headers = {\n",
|
||
" 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:123.0) Gecko/20100101 Firefox/123.0',\n",
|
||
" }\n",
|
||
" \n",
|
||
" response = requests.get(f'https://www.rightmove.co.uk/typeAhead/uknostreet/{name}', headers=headers)\n",
|
||
" return response.json()"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "d185e013-9beb-4e57-9d8b-a830e14339c3",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": []
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "cabe5d3f-ad0f-49c2-9fbf-686539a05bd1",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 30,
|
||
"id": "ba500fe6-fb18-466e-a697-403d28181674",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": []
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 46,
|
||
"id": "56c49b50-ff31-4785-9088-45ff5a39545e",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"{'Greenwich': 'REGION^61226', 'Hillingdon': 'REGION^93959', 'Ealing': 'REGION^93947', 'Richmond upon Thames': 'REGION^61415', 'Sutton': 'REGION^93974', 'Wandsworth': 'REGION^93977', 'Camden': 'REGION^93941', 'Enfield': 'REGION^93950', 'Croydon': 'REGION^93944', 'Hackney': 'REGION^93953', 'Kingston upon Thames': 'REGION^93968', 'Kensington and Chelsea': 'REGION^61229', 'Bromley': 'REGION^93938', 'Brent': 'REGION^93935', 'Waltham Forest': 'REGION^61232', 'Southwark': 'REGION^61518', 'Harrow': 'REGION^93956', 'Lewisham': 'REGION^61413', 'Barnet': 'REGION^93929', 'Islington': 'REGION^93965', 'Haringey': 'REGION^61227', 'Lambeth': 'REGION^93971', 'Westminster': '', 'Tower Hamlets': 'REGION^61417', 'Havering': 'REGION^61228', 'Barking and Dagenham': 'REGION^61400', 'Hammersmith and Fulham': 'REGION^61407', 'Bexley': 'REGION^93932', 'Redbridge': 'REGION^61537', 'Newham': 'REGION^61231', 'Merton': 'REGION^61414', 'Hounslow': 'REGION^93962'}\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"{\n",
|
||
" \"City of London\": \"REGION^61224\",\n",
|
||
" \"Greenwich\": \"REGION^61226\",\n",
|
||
" \"Hillingdon\": \"REGION^93959\",\n",
|
||
" \"Ealing\": \"REGION^93947\",\n",
|
||
" \"Richmond upon Thames\": \"REGION^61415\",\n",
|
||
" \"Sutton\": \"REGION^93974\",\n",
|
||
" \"Wandsworth\": \"REGION^93977\",\n",
|
||
" \"Camden\": \"REGION^93941\",\n",
|
||
" \"Enfield\": \"REGION^93950\",\n",
|
||
" \"Croydon\": \"REGION^93944\",\n",
|
||
" \"Hackney\": \"REGION^93953\",\n",
|
||
" \"Kingston upon Thames\": \"REGION^93968\",\n",
|
||
" \"Kensington and Chelsea\": \"REGION^61229\",\n",
|
||
" \"Bromley\": \"REGION^93938\",\n",
|
||
" \"Brent\": \"REGION^93935\",\n",
|
||
" \"Waltham Forest\": \"REGION^61232\",\n",
|
||
" \"Southwark\": \"REGION^61518\",\n",
|
||
" \"Harrow\": \"REGION^93956\",\n",
|
||
" \"Lewisham\": \"REGION^61413\",\n",
|
||
" \"Barnet\": \"REGION^93929\",\n",
|
||
" \"Islington\": \"REGION^93965\",\n",
|
||
" \"Haringey\": \"REGION^61227\",\n",
|
||
" \"Lambeth\": \"REGION^93971\",\n",
|
||
" \"Westminster\": \"REGION^93980\",\n",
|
||
" \"Tower Hamlets\": \"REGION^61417\",\n",
|
||
" \"Havering\": \"REGION^61228\",\n",
|
||
" \"Barking and Dagenham\": \"REGION^61400\",\n",
|
||
" \"Hammersmith and Fulham\": \"REGION^61407\",\n",
|
||
" \"Bexley\": \"REGION^93932\",\n",
|
||
" \"Redbridge\": \"REGION^61537\",\n",
|
||
" \"Newham\": \"REGION^61231\",\n",
|
||
" \"Merton\": \"REGION^61414\",\n",
|
||
" \"Hounslow\": \"REGION^93962\",\n",
|
||
"}\n",
|
||
"\n",
|
||
"bor_to_locid = {}\n",
|
||
"\n",
|
||
"for borough in boroughs:\n",
|
||
" bor_to_locid[borough] = ''\n",
|
||
" d = query_loc_identifier(borough)\n",
|
||
" locs = d['typeAheadLocations']\n",
|
||
" filtered = [l for l in locs if 'Borough' in l['displayName']]\n",
|
||
" if len(filtered)>1:\n",
|
||
" print(f\"{borough} has more entries: {len(filtered)}!\")\n",
|
||
" if filtered:\n",
|
||
" bor_to_locid[borough] = filtered[0]['locationIdentifier']\n",
|
||
"\n",
|
||
"Westminster (City of)\n",
|
||
"print(bor_to_locid)"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "157e20a2-d137-45b8-802b-948fa8e04ba3",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": []
|
||
}
|
||
],
|
||
"metadata": {
|
||
"kernelspec": {
|
||
"display_name": "Python 3 (ipykernel)",
|
||
"language": "python",
|
||
"name": "python3"
|
||
},
|
||
"language_info": {
|
||
"codemirror_mode": {
|
||
"name": "ipython",
|
||
"version": 3
|
||
},
|
||
"file_extension": ".py",
|
||
"mimetype": "text/x-python",
|
||
"name": "python",
|
||
"nbconvert_exporter": "python",
|
||
"pygments_lexer": "ipython3",
|
||
"version": "3.12.0"
|
||
}
|
||
},
|
||
"nbformat": 4,
|
||
"nbformat_minor": 5
|
||
}
|