INTRODUCTION

About

This work is a personal study on Seattle's housing market and it is not for commercial use. Unfortunately, the data is limited to 2014-2015 and extension of the work to today's [crazy] market is not practical. Data exploration and ML modeling is captured in multiple "jupyter notebooks" in this repository. The problem is initially a regression modeling exercise. Simple linear model and tree-based algorithms was used. Although Random Forest (like always) offered a significant improvement to the model, for the purpose of statistical inference, improved (Ridge) linear model was chosen for this study.

Acknowledgement: The original dataset belongs to Kaggle: House Sales in King County, USA provided to redict house price using regression.

Data

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

data = pd.read_csv("clean_data.csv")

data.drop(['Unnamed: 0'], axis=1, inplace=True)

data.head()

<style> .dataframe thead tr:only-child th { text-align: right; }

.dataframe thead th {
    text-align: left;
}

.dataframe tbody tr th {
    vertical-align: top;
}

</style>

	price	bedrooms	bathrooms	sqft_living	sqft_lot	floors	waterfront	view	condition	grade	...	sqft_basement	yr_built	yr_renovated	zipcode	lat	long	sqft_living15	sqft_lot15	year	month
0	12.309982	3	1.00	1180	8.639411	1.0	0	0	3	7	...	0	1955	0	98178	47.5112	-122.257	1340	8.639411	2014	10
1	13.195614	3	2.25	2570	8.887653	2.0	0	0	3	7	...	400	1951	1	98125	47.7210	-122.319	1690	8.941022	2014	12
2	12.100712	2	1.00	770	9.210340	1.0	0	0	3	6	...	0	1933	0	98028	47.7379	-122.233	2720	8.994917	2015	2
3	13.311329	4	3.00	1960	8.517193	1.0	0	0	5	7	...	910	1965	0	98136	47.5208	-122.393	1360	8.517193	2014	12
4	13.142166	3	2.00	1680	8.997147	1.0	0	0	3	8	...	0	1987	0	98074	47.6168	-122.045	1800	8.923058	2015	2

5 rows × 21 columns

import plotly
import plotly.graph_objs as go

plotly.tools.set_credentials_file(username='----', api_key='----')
mapbox_access_token = '-----'

data_ = [
    go.Scattermapbox(
        lat=list(data.lat),
        lon=list(data.long),
        mode='markers',
        marker=dict(size=2),
        text=[''],
    )
]

layout = go.Layout(
    autosize=True,
    hovermode='closest',
    mapbox=dict(
        accesstoken=mapbox_access_token,
        bearing=0,
        center=dict(
            lat=47.5,
            lon=-122
        ),
        pitch=0,
        zoom=5
    ),
)

fig = dict(data=data_, layout=layout);
plotly.plotly.iplot(fig)

Fin!