Ford Used Car Price Prediction

This project focuses on building a machine learning-powered web dashboard to assist Ford used car dealers in predicting and managing car prices. Using historical data, three models—Linear Regression, Gradient Boosting Regressor, and Random Forest—were trained and evaluated. The Random Forest Regressor delivered the most accurate results and was selected for deployment. The predicted prices were integrated into a dynamic PHP web application, allowing users to filter listings by model, year, and fuel type.

Tools Used: Python, Scikit-learn, Google Colab, MySQL, PHP, AJAX.

🔗 Launch Dealer Dashboard

Linear Regression Random Forest Gradient Boosting

🚗 Ford Used Car Price Prediction & Dealer Dashboard

Objective:
Build a machine learning-powered system to help Ford dealers search, compare, and price used cars more accurately.

---

🛠 Tools Used:

• Python (Pandas, Scikit-learn, Matplotlib)
• Google Colab (Model training and data prep)
• MySQL (Database for storing enriched dataset)
• PHP + AJAX + jQuery (Web dashboard)
• Hostinger (Live deployment)
• HTML/CSS (User interface)

---

🧠 Machine Learning Models Trained:

• Linear Regression
• Gradient Boosting Regressor
• ✅ Random Forest Regressor (Best Performer)

Random Forest showed the best R² score and lowest RMSE for predicting used car prices.

---

🧾 Features Used to Predict Price:

The Random Forest model predicts the price based on the following key features:

| Feature | Why It Matters

| model | Different models have different base values
| year | Newer cars typically cost more
| mileage | Lower mileage suggests better condition
| fuelType | Fuel preference affects buyer demand
| transmission | Automatic/manual affects driving preference
| engineSize | Larger engines affect performance and fuel economy
| tax | Higher tax = more ownership cost
| mpg | Better fuel economy increases value

---

🔧 Project Workflow:

1. Cleaned and explored Ford used car dataset
2. Trained and compared 3 regression models
3. Selected best model (Random Forest) and added predicted prices
4. Stored results in MySQL database on Hostinger
5. Built a PHP dashboard with dynamic dropdowns:
   • Model → Year → Fuel Type
6. Displayed full car details + predicted prices in a clean table

---

💼 Business Value for Ford Dealers:

• 💰 Smarter pricing using ML-predicted price
• 🔍 Faster filtering by model, year, and fuel type
• 📊 Data-backed decisions for sales and inventory
• 🌐 Online access from any device – no software needed

---

✅ Future Enhancements:

• Add image uploads or PDF quotes
• Dealer login and pricing override
• Export results as Excel/CSV

Importing Dataset

from google.colab import files
uploaded = files.upload()

Upload widget is only available when the cell has been executed in the current browser session. Please rerun this cell to enable.

Saving ford.csv to ford.csv

import pandas as pd
df = pd.read_csv('ford.csv')
df.head()

	model	year	price	transmission	mileage	fuelType	tax	mpg	engineSize
0	Fiesta	2017	12000	Automatic	15944	Petrol	150	57.7	1.0
1	Focus	2018	14000	Manual	9083	Petrol	150	57.7	1.0
2	Focus	2017	13000	Manual	12456	Petrol	150	57.7	1.0
3	Fiesta	2019	17500	Manual	10460	Petrol	145	40.3	1.5
4	Fiesta	2019	16500	Automatic	1482	Petrol	145	48.7	1.0

# Check for missing values
print(df.isnull().sum())

# Check data types
print(df.dtypes)

model           0
year            0
price           0
transmission    0
mileage         0
fuelType        0
tax             0
mpg             0
engineSize      0
dtype: int64
model            object
year              int64
price             int64
transmission     object
mileage           int64
fuelType         object
tax               int64
mpg             float64
engineSize      float64
dtype: object

df_encoded = pd.get_dummies(df, columns=['model', 'transmission', 'fuelType'], drop_first=True)
df_encoded.shape

(17966, 35)

X = df_encoded.drop('price', axis=1)
y = df_encoded['price']

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)
print("Training set size:", X_train.shape)
print("Test set size:", X_test.shape)

Training set size: (14372, 34)
Test set size: (3594, 34)

Linear Regression Model

from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
import numpy as np

lr_model = LinearRegression()
lr_model.fit(X_train, y_train)


y_pred = lr_model.predict(X_test)


mae = mean_absolute_error(y_test, y_pred)
rmse = np.sqrt(mean_squared_error(y_test, y_pred))
r2 = r2_score(y_test, y_pred)

print("Linear Regression Performance:")
print(f"MAE: {mae:.2f}")
print(f"RMSE: {rmse:.2f}")
print(f"R² Score: {r2:.2f}")

Linear Regression Performance:
MAE: 1374.81
RMSE: 1858.76
R² Score: 0.85

import io

df_original = pd.read_csv(io.BytesIO(uploaded['ford.csv']))

df_original['lr_predictedprice'] = lr_model.predict(X)


df_original.head()

	model	year	price	transmission	mileage	fuelType	tax	mpg	engineSize	lr_predictedprice
0	Fiesta	2017	12000	Automatic	15944	Petrol	150	57.7	1.0	11115.468735
1	Focus	2018	14000	Manual	9083	Petrol	150	57.7	1.0	14286.297386
2	Focus	2017	13000	Manual	12456	Petrol	150	57.7	1.0	12994.669108
3	Fiesta	2019	17500	Manual	10460	Petrol	145	40.3	1.5	15832.708316
4	Fiesta	2019	16500	Automatic	1482	Petrol	145	48.7	1.0	14900.975985

import matplotlib.pyplot as plt

sample_size = 100
plt.figure(figsize=(14, 6))

plt.plot(df_original['price'][:sample_size].values, label='Actual Price', marker='o')

plt.plot(df_original['lr_predictedprice'][:sample_size].values, label='Predicted Price', marker='x')

plt.title('Actual vs Predicted Car Prices (First 100 Cars)')
plt.xlabel('Car Index')
plt.ylabel('Price')
plt.legend()
plt.grid(True)
plt.tight_layout()
plt.show()

Random Forest

from sklearn.ensemble import RandomForestRegressor

rf_model = RandomForestRegressor(n_estimators=100, random_state=42)
rf_model.fit(X_train, y_train)

y_pred_rf = rf_model.predict(X_test)

mae_rf = mean_absolute_error(y_test, y_pred_rf)
rmse_rf = np.sqrt(mean_squared_error(y_test, y_pred_rf))
r2_rf = r2_score(y_test, y_pred_rf)

print("Random Forest Regressor Performance:")
print(f"MAE: {mae_rf:.2f}")
print(f"RMSE: {rmse_rf:.2f}")
print(f"R² Score: {r2_rf:.2f}")

Random Forest Regressor Performance:
MAE: 863.83
RMSE: 1258.82
R² Score: 0.93

import io

df_original_rf = pd.read_csv(io.BytesIO(uploaded['ford.csv']))

df_original_rf['rf_model_predictedprice'] = rf_model.predict(X)

df_original_rf.head()

	model	year	price	transmission	mileage	fuelType	tax	mpg	engineSize	rf_model_predictedprice
0	Fiesta	2017	12000	Automatic	15944	Petrol	150	57.7	1.0	10407.50
1	Focus	2018	14000	Manual	9083	Petrol	150	57.7	1.0	13907.83
2	Focus	2017	13000	Manual	12456	Petrol	150	57.7	1.0	12870.18
3	Fiesta	2019	17500	Manual	10460	Petrol	145	40.3	1.5	18097.42
4	Fiesta	2019	16500	Automatic	1482	Petrol	145	48.7	1.0	16264.99

import matplotlib.pyplot as plt


sample_size = 100
plt.figure(figsize=(14, 6))

plt.plot(df_original_rf['price'][:sample_size].values, label='Actual Price', marker='o')

plt.plot(df_original_rf['rf_model_predictedprice'][:sample_size].values, label='Predicted Price', marker='x')

plt.title('Actual vs Predicted Car Prices (First 100 Cars)')
plt.xlabel('Car Index')
plt.ylabel('Price')
plt.legend()
plt.grid(True)
plt.tight_layout()
plt.show()

Gradient Boosting

from sklearn.ensemble import GradientBoostingRegressor


gb_model = GradientBoostingRegressor(n_estimators=100, learning_rate=0.1, random_state=42)
gb_model.fit(X_train, y_train)


y_pred_gb = gb_model.predict(X_test)

mae_gb = mean_absolute_error(y_test, y_pred_gb)
rmse_gb = np.sqrt(mean_squared_error(y_test, y_pred_gb))
r2_gb = r2_score(y_test, y_pred_gb)

print("Gradient Boosting Regressor Performance:")
print(f"MAE: {mae_gb:.2f}")
print(f"RMSE: {rmse_gb:.2f}")
print(f"R² Score: {r2_gb:.2f}")

Gradient Boosting Regressor Performance:
MAE: 975.93
RMSE: 1381.04
R² Score: 0.91

import io

df_original = pd.read_csv(io.BytesIO(uploaded['ford.csv']))

df_original['gb_model_predictedprice'] = gb_model.predict(X)

df_original.head()

	model	year	price	transmission	mileage	fuelType	tax	mpg	engineSize	gb_model_predictedprice
0	Fiesta	2017	12000	Automatic	15944	Petrol	150	57.7	1.0	10786.833313
1	Focus	2018	14000	Manual	9083	Petrol	150	57.7	1.0	13617.000502
2	Focus	2017	13000	Manual	12456	Petrol	150	57.7	1.0	11711.747945
3	Fiesta	2019	17500	Manual	10460	Petrol	145	40.3	1.5	16939.455232
4	Fiesta	2019	16500	Automatic	1482	Petrol	145	48.7	1.0	16323.893499

import matplotlib.pyplot as plt


sample_size = 100
plt.figure(figsize=(14, 6))


plt.plot(df_original['price'][:sample_size].values, label='Actual Price', marker='o')


plt.plot(df_original['gb_model_predictedprice'][:sample_size].values, label='Predicted Price', marker='x')

plt.title('Actual vs Predicted Car Prices (First 100 Cars)')
plt.xlabel('Car Index')
plt.ylabel('Price')
plt.legend()
plt.grid(True)
plt.tight_layout()
plt.show()

Saving predicted values into csv file

from google.colab import files

output_path = "ford_with_predicted_prices_rf.csv"
df_original_rf.to_csv(output_path, index=False)

files.download(output_path)

Converting CSV file to sql file

df_with_predictions = pd.read_csv("ford_with_predicted_prices_rf.csv")

table_name = "cars"
columns = df_with_predictions.columns

# Create table schema
create_table_sql = f"CREATE TABLE {table_name} (\n"
for col in columns:
    if df_with_predictions[col].dtype == 'int64':
        col_type = "INT"
    elif df_with_predictions[col].dtype == 'float64':
        col_type = "FLOAT"
    else:
        col_type = "VARCHAR(255)"
    create_table_sql += f"  `{col}` {col_type},\n"
create_table_sql = create_table_sql.rstrip(",\n") + "\n);\n"

# Generate INSERT statements
insert_statements = ""
for _, row in df_with_predictions.iterrows():
    escaped_values = []
    for val in row:
        if isinstance(val, str):
            escaped_val = val.replace("'", "''")  # Escape single quotes
            escaped_values.append(f"'{escaped_val}'")
        else:
            escaped_values.append(f"'{val}'")
    values = ', '.join(escaped_values)
    insert_statements += f"INSERT INTO {table_name} VALUES ({values});\n"

# Combine all SQL
full_sql = create_table_sql + "\n" + insert_statements

# Save to .sql file
sql_path = "car_prices.sql"
with open(sql_path, "w", encoding='utf-8') as f:
    f.write(full_sql)

sql_path
files.download(sql_path)