SPB Git

spb/vquant Public MIT

VibeQuant — AI-powered institutional-grade financial intelligence platform.

TypeScript 84.3% Python 11.7% JavaScript 1.6% CSS 1.5% HTML 0.7%
6.7 KB

# Python Quantitative Analysis

VibeQuant includes a powerful Python engine for quantitative financial analysis. All Python code executes in an isolated environment with 17 scientific libraries.

# Built-in Analysis Models

# 1. Monte Carlo Simulation

Simulate future price paths using geometric Brownian motion.

Parameters:

Parameter Type Default Description
symbol string required Stock ticker symbol
num_simulations int 10,000 Number of simulation paths
time_horizon int 252 Trading days to simulate
initial_investment float 10,000 Starting portfolio value

Example prompt:

text
"Run a Monte Carlo simulation on NVDA with 50,000 paths over 500 trading days"

Output: Distribution statistics, probability of profit, VaR estimates, simulation chart.


# 2. Options Pricing (Black-Scholes)

Price European options with full Greeks calculation.

Parameters:

Parameter Type Description
symbol string Underlying asset
strike_price float Strike price
expiry_days int Days to expiration
risk_free_rate float Risk-free rate (default: 0.05)
option_type string "call" or "put"

Output: Option price, delta, gamma, theta, vega, rho, put-call parity check.

Example prompt:

text
"Price a call option on AAPL with strike $200, 30 days to expiry"

# 3. GARCH Volatility Model

Estimate and forecast volatility using GARCH(1,1).

Parameters:

Parameter Type Description
symbol string Stock ticker
data_period int Historical days (default: 756)
forecast_horizon int Days to forecast (default: 30)

Output: GARCH parameters (omega, alpha, beta), historical volatility, forecast, interpretation.


# 4. Value at Risk (VaR)

Calculate VaR using three methods: Historical, Parametric, and Monte Carlo.

Parameters:

Parameter Type Default Description
symbol string required Stock ticker
portfolio_value float 100,000 Portfolio value
confidence_levels float[] [0.90, 0.95, 0.99] Confidence levels
time_horizon int 1 Holding period (days)
num_simulations int 10,000 MC simulations

Output: VaR by method and confidence level, CVaR (Expected Shortfall), distribution statistics.


# 5. Portfolio Optimization

Mean-variance optimization using Modern Portfolio Theory.

Parameters:

Parameter Type Default Description
symbols string[] required Portfolio tickers
risk_free_rate float 0.05 Risk-free rate
min_weight float 0.0 Minimum allocation per asset
max_weight float 1.0 Maximum allocation per asset
generate_frontier bool false Generate efficient frontier

Example prompt:

text
"Optimize a portfolio of AAPL, MSFT, GOOGL, AMZN, NVDA for maximum Sharpe ratio"

Output: Optimal weights (max Sharpe, min variance, equal weight), efficient frontier, correlation matrix.


# 6. Risk Metrics Analysis

Comprehensive risk-adjusted performance metrics.

Parameters:

Parameter Type Default Description
symbol string required Stock ticker
benchmark_symbol string SPY Benchmark
risk_free_rate float 0.05 Risk-free rate

Output:

  • Return metrics: Annualized return, volatility, excess return
  • Risk-adjusted: Sharpe, Sortino, Calmar, Information, Treynor ratios
  • Market: Beta, Alpha, Correlation, R-squared, Tracking error
  • Drawdown: Max drawdown, recovery days
  • Trading: Win rate, profit factor, average win/loss

# 7. Chart Generation (create_plot)

Create publication-quality charts.

Plot types: line, bar, scatter, histogram, candlestick, area, pie, heatmap

Parameters:

Parameter Type Description
plot_type string Chart type
data object Data to plot
title string Chart title
xlabel / ylabel string Axis labels
color string/string[] Colors
figsize [w, h] Figure size
theme string "default", "dark", "colorful"

# 8. Custom Python Code

Execute arbitrary Python code with access to all 17 libraries and FMP data.

Example prompt:

text
"Write Python code to:
1. Fetch 2 years of daily prices for AAPL and MSFT
2. Calculate 30-day rolling correlation
3. Plot correlation over time with a heatmap of monthly averages"

The AI will generate and execute Python code, returning charts and output.

Available in custom code:

  • All 17 scientific libraries
  • fmp_api_key for fetching financial data
  • Matplotlib figure generation (saved as PNG)
  • Print output captured and returned

# Python Libraries Reference

Library Import Key Functions
numpy import numpy as np Arrays, linear algebra, random
pandas import pandas as pd DataFrames, time series, IO
scipy from scipy import stats, optimize Statistics, optimization
scikit-learn from sklearn import ... ML models, preprocessing
statsmodels import statsmodels.api as sm OLS, ARIMA, tests
matplotlib import matplotlib.pyplot as plt Plotting
seaborn import seaborn as sns Statistical plots
plotly import plotly.graph_objects as go Interactive charts
arch from arch import arch_model GARCH, EGARCH
cvxpy import cvxpy as cp Convex optimization
yfinance import yfinance as yf Market data
ta import ta 40+ technical indicators
quantstats import quantstats as qs Portfolio analytics
beautifulsoup4 from bs4 import BeautifulSoup HTML parsing
requests import requests HTTP client
lxml from lxml import etree XML parsing
Pillow from PIL import Image Image processing