Title: "Exploring Determinants of House Prices: A Comprehensive Analysis"

Authors:

Research questions

How does the age of a house affect its price in the housing market?

Hypothesis 1: Age and Renovation Impact on Price: This hypothesis could be that the age of the house
and whether it has been renovated or not could impact its price. Newer or recently renovated houses
might command higher prices due to their updated features and modern amenities compared to older,
unrenovated properties.

How does the price of houses vary across different cities or neighborhoods?

Hypothesis 2: Price Variation Based on Location: We might hypothesize that the price of houses varies
significantly based on their location (i.e., city or neighborhood). This is because different cities or
neighborhoods may have varying levels of amenities, infrastructure, and desirability, all of which can
influence property prices.

Motivation and background

Motivation for Research Question 1 (Age and Renovation Impact on Price):

Market Dynamics: Understanding how the age and renovation status of a house influence its price is
essential for both buyers and sellers in the real estate market. Buyers seek to invest in properties that
offer the best value for their money, while sellers aim to maximize their returns on investment.

Consumer Preferences: Researching the specific features or improvements that contribute to increased
property value helps to align seller renovations with buyer preferences. This knowledge can guide sellers
in making informed decisions about which renovations to undertake to maximize their property's resale
value.

Investment Strategies: Investors in the real estate market also rely on insights into how age and
renovation impact property prices to make strategic investment decisions. Understanding the return on
investment (ROI) associated with renovations versus purchasing newer properties can inform investment
strategies.

Motivation for Research Question 2 (Price Variation Based on Location):

Spatial Economics: Real estate prices vary significantly across different locations due to factors such as
amenities, infrastructure, proximity to employment centers, and quality of schools. Understanding these
spatial variations is crucial for policymakers, urban planners, and real estate professionals.

Housing Affordability: Researching the price variation based on location sheds light on housing
affordability issues within different regions. It helps policymakers identify areas where housing
affordability is a concern and develop targeted policies to address the needs of residents.

Market Segmentation: For real estate professionals, understanding the price variation based on location
enables them to segment the market effectively. They can tailor marketing strategies and pricing
decisions based on the unique characteristics and demands of each location, maximizing their
competitiveness and profitability.

Dataset

The dataset that I provided is from Kaggle website it contain information about real estate properties,
likely collected for analysis or modeling purposes. Here's a description of the variables in the dataset:

1. Date (Variable Type: Qualitative)

Description: Date of sale of the property.

Level of Measurement: Nominal

Unit of Measurement: Date (in the format YYYY-MM-DD)

2. Price (Variable Type: Quantitative)

Description: Sale price of the property.

Level of Measurement: Continuous

Unit of Measurement: Currency (presumably in the local currency, such as USD)

Bedrooms (Variable Type: Quantitative)

3. Description: Number of bedrooms in the property.

Level of Measurement: Discrete

Unit of Measurement: Count

4. Bathrooms (Variable Type: Quantitative)

Description: Number of bathrooms in the property, including fractional bathrooms.

Level of Measurement: Discrete

Unit of Measurement: Count (could be fractional)

5. Sqft_living (Variable Type: Quantitative)

Description: Square footage of living space in the property.

Level of Measurement: Continuous

Unit of Measurement: Square Feet

6. Sqft_lot (Variable Type: Quantitative)

Description: Square footage of the lot (land) on which the property is situated.

Level of Measurement: Continuous

Unit of Measurement: Square Feet

7. Floors (Variable Type: Quantitative)

Description: Number of floors in the property.

Level of Measurement: Discrete

Unit of Measurement: Count

8. Waterfront (Variable Type: Qualitative)

Description: Indicates whether the property has a waterfront view or not.

Level of Measurement: Nominal (binary)

Unit of Measurement: Binary (0 for no waterfront, 1 for waterfront)

9. View (Variable Type: Quantitative)

Description: Level of view quality from the property (0-4).

Level of Measurement: Ordinal

Unit of Measurement: Scale (0 to 4)

10. Condition (Variable Type: Quantitative)

Description: Overall condition of the property (1-5, with 5 being the best).

Level of Measurement: Ordinal

Unit of Measurement: Scale (1 to 5)

11. Sqft_above (Variable Type: Quantitative)

Description: Square footage of living space above ground level.

Level of Measurement: Continuous

Unit of Measurement: Square Feet

12. Sqft_basement (Variable Type: Quantitative)

Description: Square footage of the basement in the property.

Level of Measurement: Continuous

Unit of Measurement: Square Feet

13. Yr_built (Variable Type: Quantitative)

Description: Year the property was built.

Level of Measurement: Discrete

Unit of Measurement: Year

14. Yr_renovated (Variable Type: Quantitative)

Description: Year the property was last renovated. (0 indicates no renovation)

Level of Measurement: Discrete

Unit of Measurement: Year

15. Street, City, Statezip, Country (Variable Type: Qualitative)

Description: Address information for the property, including street name, city, state zip code, and
country.

Level of Measurement: Nominal

Unit of Measurement: Textual

Methods used

Hypothesis one

Importing necessary libraries:

Similar to Code 1, this step imports essential libraries required for data analysis, including pandas,
numpy, matplotlib, seaborn, and scikit-learn.

Loading the dataset:

Uses pd.read_csv() to load the dataset from a CSV file into a pandas DataFrame called data.

Exploring the dataset:

data.head(): Displays the first few rows of the dataset to get an initial overview.

data.info(): Provides summary information about the dataset, including data types and missing values.

Exploratory Data Analysis (EDA):

sns.scatterplot(): Visualizes the relationship between the year built, renovation year, and price using a
scatter plot with the hue representing renovation year.

Linear Regression:

train_test_split(): Splits the dataset into training and testing sets.

LinearRegression(): Initializes a linear regression model.

model.fit(): Fits the linear regression model to the training data.

model.predict(): Predicts house prices using the trained model on the testing set.

mean_squared_error(): Calculates the Mean Squared Error (MSE) to evaluate the model's performance.

plt.scatter(): Visualizes the predicted vs. actual house prices using a scatter plot.

Why these methods have been used:

Exploratory Data Analysis (EDA):

EDA methods such as displaying dataset overview (head()), checking summary information (info()), and
computing summary statistics (describe()) are used to understand the dataset's structure, contents, and
distribution of variables. Visualization methods like scatter plots and box plots help identify patterns,
trends, and relationships within the data.

Linear Regression:

Linear regression is used to analyze the relationship between independent variables (e.g., year built,
renovation year) and the dependent variable (price). It helps in building a predictive model to estimate
house prices based on other property features. Methods like train_test_split() and
mean_squared_error() are employed to evaluate the performance of the regression model.

Hypothesis two

Importing necessary libraries:

This step imports essential libraries like pandas, numpy, matplotlib, seaborn, and scikit-learn. These
libraries provide tools for data manipulation, visualization, and machine learning model building.

Loading the dataset:

The pd.read_csv() function is used to load the dataset from a CSV file into a pandas DataFrame. This is
the initial step to access and analyze the dataset.

Exploratory Data Analysis (EDA):

print(df.head()): Displays the first few rows of the dataset to understand its structure and contents.

print(df.info()): Provides information about the dataset, including data types and missing values.

print(df.describe()): Computes summary statistics to understand the distribution of numerical variables.

print(df.isnull().sum()): Checks for missing values in the dataset.

Visualization methods like box plots (sns.boxplot()) are used to visualize the distribution of house prices
across different cities/neighborhoods.

Linear Regression:

pd.get_dummies(): Encodes categorical variables using one-hot encoding to prepare them for linear
regression analysis.

train_test_split(): Splits the dataset into training and testing sets for model evaluation.

LinearRegression(): Initializes a linear regression model.

model.fit(): Fits the linear regression model to the training data.

model.predict(): Predicts house prices using the trained model on the testing set.

mean_squared_error(): Calculates the Mean Squared Error (MSE) to evaluate the model's performance.

plt.scatter(): Visualizes the predicted vs. actual house prices using a scatter plot.

Result and conclusion

Hypothesis 1:

The analysis conducted in the provided code examines the relationship between house prices and
variables related to the year built and renovation year. Through exploratory data analysis and linear
regression modeling, the study aims to understand how these factors influence housing prices. However,
the scatter plot visualization demonstrates a wide range of house prices across different years built and
renovation years, but it fails to provide clear insights due to data overlap. Additionally, the linear
regression model yields a high Mean Squared Error (MSE), indicating significant discrepancies between
actual and predicted house prices. This suggests that the model may not effectively capture the
variability in house prices based solely on year built and renovation year. Consequently, it is apparent
that the dataset may lack crucial variables, such as location-specific factors and property attributes,
which could better explain house price variations. Further analysis with more advanced modeling
techniques and inclusion of additional relevant variables may be necessary to improve predictive
accuracy and better understand the determinants of housing prices.

In conclusion, while the analysis sheds some light on the relationship between year built, renovation
year, and house prices, it highlights the limitations of linear regression in capturing complex nonlinear
relationships. The implications underscore the importance of thorough feature engineering, model
selection, and validation procedures in real estate valuation tasks. To enhance model performance and
predictive accuracy, future research should consider incorporating a more comprehensive set of
variables and exploring alternative modeling approaches, such as tree-based methods or neural
networks. By addressing these limitations, researchers can develop more robust predictive models that
provide valuable insights into housing market dynamics and facilitate informed decision-making for
buyers, sellers, and real estate professionals.
Hypothesis 2:

The analysis conducted in the provided code delves into understanding the connection between house
prices and location, particularly focusing on various cities or neighborhoods. Through exploratory data
analysis (EDA), the distribution of house prices across different locations is visualized using boxplots,
shedding light on the variations in property values among different cities. Subsequently, a linear
regression model is employed to quantify the relationship between house prices and location variables,
specifically by encoding categorical variables like city and attempting to predict house prices based on
these attributes. However, the relatively high mean squared error (MSE) obtained from the linear
regression analysis indicates that location alone may not suffice to accurately predict house prices,
suggesting the influence of other significant factors beyond location.

The outcomes of the analysis underscore the importance of considering a broader spectrum of variables
beyond location to better comprehend the determinants of house prices. While location undoubtedly
plays a pivotal role, factors such as property characteristics, neighborhood amenities, economic
indicators, and market trends are crucial contributors to house price variations. Future research
endeavors could explore additional quantitative variables and qualitative factors such as neighborhood
desirability and safety to develop more comprehensive predictive models for house prices, enhancing
the explanatory power and predictive accuracy of such models. Ultimately, the analysis highlights the
complexity of determining house prices and emphasizes the necessity of a holistic approach
encompassing various factors to gain deeper insights into real estate market dynamics.