ai.py

# -*- coding: utf-8 -*-
"""Copy of Ai.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1vUpFvkYZHVz6r6jYkIMHiw63Gwj84xvy
"""

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

import pandas as pd

trades = pd.read_csv("historical_data.csv")
fear = pd.read_csv("fear_greed_index (1).csv")

print(trades.head())
print(fear.head())

trades.info()
fear.info()

trades = trades.drop_duplicates()
fear = fear.drop_duplicates()

trades['Date'] = pd.to_datetime(trades['Timestamp IST'], format='%d-%m-%Y %H:%M')
fear['Date'] = pd.to_datetime(fear['date'])

merged = pd.merge(
    trades,
    fear,
    on='Date',
    how='left'
)

merged.groupby('classification')['Closed PnL'].mean()

merged.groupby('classification')['Size USD'].mean()

merged.groupby('classification')['Size USD'].mean()

merged['win'] = merged['Closed PnL'] > 0

merged.groupby('classification')['win'].mean()

import matplotlib.pyplot as plt

# Sort the merged DataFrame by 'Date' to ensure the line plot is sequential
merged_sorted = merged.sort_values(by='Date')

plt.figure(figsize=(15, 7))
plt.plot(merged_sorted['Date'], merged_sorted['Closed PnL'], label='Closed PnL Over Time')
plt.title('Closed PnL Trend Over Time')
plt.xlabel('Date')
plt.ylabel('Closed PnL')
plt.grid(True)
plt.legend()
plt.tight_layout()
plt.show()

import matplotlib.pyplot as plt
import seaborn as sns

# Prepare data: Sort by date and calculate cumulative profit
plot_df = merged.sort_values('Date').dropna(subset=['value'])
plot_df['Cumulative PnL'] = plot_df['Closed PnL'].cumsum()

fig, ax1 = plt.subplots(figsize=(16, 8))

# Plot 1: Fear & Greed Index (Area Plot)
ax1.fill_between(plot_df['Date'], plot_df['value'], color='gray', alpha=0.2, label='Fear & Greed Index')
ax1.plot(plot_df['Date'], plot_df['value'], color='gray', alpha=0.5, linewidth=1)
ax1.set_ylabel('Fear & Greed Index Value', color='gray', fontsize=12)
ax1.tick_params(axis='y', labelcolor='gray')
ax1.set_ylim(0, 100) # Fear & Greed is always 0-100

# Create a second y-axis for Cumulative PnL
ax2 = ax1.twinx()
ax2.plot(plot_df['Date'], plot_df['Cumulative PnL'], color='#1f77b4', linewidth=2.5, label='Cumulative Profit (USD)')
ax2.set_ylabel('Cumulative Profit (USD)', color='#1f77b4', fontsize=12)
ax2.tick_params(axis='y', labelcolor='#1f77b4')

# Add horizontal line at 0 for Profit
ax2.axhline(0, color='red', linestyle='--', linewidth=1, alpha=0.5)

# Styling
plt.title('Historical Performance vs. Market Sentiment', fontsize=16, fontweight='bold')
ax1.set_xlabel('Date', fontsize=12)
ax1.grid(True, which='both', linestyle=':', alpha=0.5)

# Add legend
lines1, labels1 = ax1.get_legend_handles_labels()
lines2, labels2 = ax2.get_legend_handles_labels()
ax1.legend(lines1 + lines2, labels1 + labels2, loc='upper left')

plt.tight_layout()
plt.show()

import matplotlib.pyplot as plt
import pandas as pd

# Prepare sequential data
df_plot = merged.sort_values('Date').dropna(subset=['value'])
df_plot['Cum Profit'] = df_plot['Closed PnL'].cumsum()
df_plot['Rolling Win Rate'] = df_plot['win'].rolling(window=10).mean()

fig, axes = plt.subplots(4, 1, figsize=(15, 20), sharex=True)
metrics = [
    ('Cum Profit', 'Cumulative Profit (USD)', '#2ecc71'),
    ('Size USD', 'Trade Size (USD)', '#3498db'),
    ('value', 'Fear & Greed Index', '#95a5a6'),
    ('Rolling Win Rate', 'Win Rate (10-Trade Rolling)', '#e74c3c')
]

for i, (col, label, color) in enumerate(metrics):
    ax = axes[i]
    # Overlay Fear & Greed Index in background of every plot for context
    ax_bg = ax.twinx()
    ax_bg.fill_between(df_plot['Date'], df_plot['value'], color='gray', alpha=0.1)
    ax_bg.set_yticks([])

    # Plot main metric
    ax.plot(df_plot['Date'], df_plot[col], color=color, linewidth=2, label=label)
    ax.set_ylabel(label, fontweight='bold')
    ax.grid(True, alpha=0.3)
    ax.legend(loc='upper left')

plt.title('Professional Performance Dashboard', y=4.1, fontsize=16)
plt.tight_layout()
plt.show()

import matplotlib.pyplot as plt
import seaborn as sns

# Prepare summary for the heatmap
# Note: 'summary' was calculated in previous steps
heatmap_data = summary.set_index('classification')[['Closed PnL', 'Size USD', 'win']]
heatmap_data.columns = ['Avg Profit', 'Avg Size', 'Win Rate']

# Create a single 2x2 Professional Grid
fig, axes = plt.subplots(2, 2, figsize=(16, 12))
sns.set_style('whitegrid')

# 1. Violin Plot for Profit Distribution
sns.violinplot(ax=axes[0,0], x='classification', y='Closed PnL', data=merged, hue='classification', palette='viridis', inner='quartile', legend=False)
axes[0,0].set_title('Profit Density Distribution (Violin)', fontweight='bold')
axes[0,0].axhline(0, color='red', linestyle='--', alpha=0.5)

# 2. Box Plot for Trade Size
sns.boxplot(ax=axes[0,1], x='classification', y='Size USD', data=merged, hue='classification', palette='mako', legend=False)
axes[0,1].set_title('Trade Size Range & Outliers (Box)', fontweight='bold')

# 3. Performance Heatmap (KPI Overview)
sns.heatmap(ax=axes[1,0], data=heatmap_data, annot=True, cmap='RdYlGn', fmt='.2f', cbar=True)
axes[1,0].set_title('Sentiment KPI Heatmap', fontweight='bold')

# 4. Strip Plot (Individual trade outcomes)
# sampling to keep the plot clean if the dataset is large
sample_data = merged.dropna(subset=['classification']).sample(min(1000, len(merged.dropna(subset=['classification']))))
sns.stripplot(ax=axes[1,1], x='classification', y='Closed PnL', data=sample_data, hue='classification', palette='rocket', jitter=True, alpha=0.4, legend=False)
axes[1,1].set_title('Individual Trade Samples (Strip Plot)', fontweight='bold')
axes[1,1].axhline(0, color='black', linestyle='-', alpha=0.2)

for ax in axes.flat:
    ax.tick_params(axis='x', rotation=30)
    ax.set_xlabel('')

plt.suptitle('Unified Sentiment Performance Dashboard', fontsize=20, fontweight='bold', y=1.02)
plt.tight_layout()
plt.show()