import warnings
import numpy as np
import pandas as pd
import plotly.express as px
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import r2_score, mean_absolute_error, mean_squared_error
warnings.filterwarnings('ignore')

df = pd.read_csv('/content/Wprld population growth rate by cities 2024.csv')

df.info()

df.describe().T

df.head()

df.isnull().sum().sort_values(ascending=False)

df['Continent'].unique()

df[df['Continent'].isnull()]

df.duplicated().any()

df.loc[df['Continent'] == 'Oceana', 'Continent'] = 'Oceania'

con = ['North America', 'Africa', 'Europe', 'Africa', 'Europe', 'Africa', 'Europe', 'Africa', 'Europe', 'Europe', 'Europe']
ind = df[df['Continent'].isnull()].index

df['Count'] = 1

fig = px.pie(df, names='Continent', values='Count')
fig.update_layout(legend_title='Continent', title={'text': 'Distribution of Continents', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()

for c in df['Continent'].unique():
  fig = px.histogram(df[df['Continent'] == c], x='Count', y='Country', color='Country', color_discrete_sequence=px.colors.qualitative.Dark24).update_yaxes(categoryorder='total ascending')
  fig.update_layout(title={'text': f'Distribution of Country in {c}', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'}, xaxis_title='Sum of Count')
  fig.show()

for n in ['Population (2023)', 'Population (2024)', 'Growth Rate']:
  fig = px.histogram(df, x=n, y="Count",
                     marginal="box",
                     hover_data=df.columns)
  fig.update_layout(title={'text': f'Distribution of {n}', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'}, yaxis_title='Sum of Count')
  fig.show()

country = df.groupby('Country', as_index=False)[['Growth Rate']].mean()
city = df.groupby('City', as_index=False).agg({'Population (2024)': 'sum', 'Growth Rate': 'mean'})

fig = px.box(df, x='Continent', y='Growth Rate', color='Continent', color_discrete_sequence=px.colors.qualitative.Dark24)
fig.update_layout(title={'text': 'Growth Rate by Continent','y':0.95,'x':0.5,'xanchor': 'center','yanchor': 'top'})
fig.show()

fig = px.bar(country.sort_values('Growth Rate', ascending=False)[:10], x='Country', y='Growth Rate', color='Country', color_discrete_sequence=px.colors.qualitative.Dark24)
fig.update_layout(title={'text': 'Top 10 Countries with Highest Average Growth Rate', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()

fig = px.bar(city.sort_values('Growth Rate', ascending=False)[:10], x='City', y='Growth Rate', color='City', color_discrete_sequence=px.colors.qualitative.Dark24)
fig.update_layout(title={'text': 'Top 10 Cities with Highest Averate Growth Rate', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()

fig = px.bar(country.sort_values('Growth Rate')[:10], x='Country', y='Growth Rate', color='Country', color_discrete_sequence=px.colors.qualitative.Dark24)
fig.update_layout(title={'text': 'Top 10 Countries with Lowest Growth Rate', 'y':0.95, 'x':0.5, 'xanchor': 'center','yanchor': 'top'})
fig.show()

fig = px.bar(city.sort_values('Population (2024)', ascending=False)[:5], x='City', y='Population (2024)', color='City', color_discrete_sequence=px.colors.qualitative.Dark24)
fig.update_layout(title={'text': 'Top 5 Cities with Highest Population (2024)', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()

fig = px.bar(city.sort_values('Population (2024)') [:5], x='City', y='Population (2024)', color='City', color_discrete_sequence=px.colors.qualitative.Dark24)
fig.update_layout(title={'text': 'Top 5 Cities with Lowest Population (2024)', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()

df_num = df.drop(columns=['Count']).select_dtypes(include=np.number)
fig = px.imshow(df_num.corr())
fig.update_layout(title={'text': 'Correlation Between Numerical Attributes', 'y':0.95, 'x':0.5, 'xanchor': 'center', 'yanchor': 'top'})
fig.show()

X = df[['Population (2023)', 'Growth Rate']]
y = df['Population (2024)']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

model = LinearRegression()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)
print(f'R2 Score: {round(r2_score(y_test, y_pred)*100,2)}%')
print(f'Mean Absolute Error: {round(mean_absolute_error(y_test, y_pred), 2)}')
print(f'Root Mean Squared Error: {round(np.sqrt(mean_squared_error(y_test, y_pred)), 2)}\n')