Advanced ArcticDB Usage for Technical Analysis with Python

ArcticDB is a powerful tool for efficiently managing and analyzing time-series financial data. When combined with technical analysis techniques, it can streamline backtesting and algorithmic trading strategies. Below, I demonstrate advanced features like deduplication, metadata management, and multi-symbol data analysis for technical indicators.

---

Why Use ArcticDB for Technical Analysis?

• High Performance: Faster read/write operations for large datasets.
• Versioning: Keep track of data versions for backtesting.
• Metadata Support: Tag datasets with useful information (e.g., source, retrieval time).
• Batch Processing: Efficiently handle multiple stock symbols.

---

Step 1: Fetch and Store Multiple Stocks with Metadata

import yfinance as yf
from arcticdb import Arctic
import pandas as pd

DB_PATH = "lmdb://./advanced_stock_db"
LIBRARY_NAME = "technical_analysis_data"

# Connect and create library
ac = Arctic(DB_PATH)
if LIBRARY_NAME not in ac.list_libraries():
    ac.create_library(LIBRARY_NAME)

lib = ac[LIBRARY_NAME]

symbols = ["AAPL", "MSFT", "GOOG", "TSLA"]

for symbol in symbols:
    ticker = yf.Ticker(symbol)
    hist_data = ticker.history(period="5y", auto_adjust=False)

    # Add metadata with data source and timestamp
    metadata = {"source": "Yahoo Finance", "retrieval_date": pd.Timestamp.now()}

    if not hist_data.empty:
        lib.write(symbol, hist_data, metadata=metadata)
        print(f"{symbol} data stored with metadata.")

---

Step 2: Batch Read and Compute Technical Indicators

import numpy as np

# Batch read multiple stock data
symbols_data = lib.batch_read(symbols)

for symbol_data in symbols_data:
    df = symbol_data.data
    symbol = symbol_data.symbol

    # Compute Technical Indicators: Bollinger Bands and MACD
    df['SMA_20'] = df['Close'].rolling(window=20).mean()
    df['Upper_Band'] = df['SMA_20'] + 2 * df['Close'].rolling(window=20).std()
    df['Lower_Band'] = df['SMA_20'] - 2 * df['Close'].rolling(window=20).std()

    # MACD calculation
    df['EMA_12'] = df['Close'].ewm(span=12, adjust=False).mean()
    df['EMA_26'] = df['Close'].ewm(span=26, adjust=False).mean()
    df['MACD'] = df['EMA_12'] - df['EMA_26']
    df['Signal_Line'] = df['MACD'].ewm(span=9, adjust=False).mean()

    # Write back updated data
    lib.write(symbol, df)
    print(f"{symbol} updated with technical indicators.")

---

Step 3: Deduplication and Versioning

# Fetch and update AAPL data to simulate updates
result = lib.read("AAPL")
existing_data = result.data

# Simulate new overlapping data
new_data = existing_data.tail(10)  # Fake overlapping update

# Combine and deduplicate
combined = pd.concat([existing_data, new_data])
deduplicated_data = combined[~combined.index.duplicated(keep='last')]

# Write as a new version
lib.write("AAPL", deduplicated_data, prune_previous_version=False)
print("AAPL data updated without deleting previous version.")

---

Step 4: Metadata and Symbol Statistics

# Fetch metadata
result = lib.read("AAPL")
print("Metadata for AAPL:", result.metadata)

# Get symbol statistics
info = lib.get_symbol_info("AAPL")
print("Symbol Info for AAPL:", info)

---

Step 5: Visualization and Analysis

import matplotlib.pyplot as plt

# Plot Bollinger Bands and Closing Prices
df = lib.read("AAPL").data
plt.figure(figsize=(12, 6))
plt.plot(df.index, df['Close'], label='Close Price')
plt.plot(df.index, df['SMA_20'], label='20-Day SMA')
plt.fill_between(df.index, df['Upper_Band'], df['Lower_Band'], color='lightgray', label='Bollinger Bands')
plt.title("AAPL Bollinger Bands")
plt.legend()
plt.show()

---

Summary of Advanced Features

Feature	Description
Metadata Handling	Store additional metadata per dataset
Batch Processing	Handle multiple symbols simultaneously
Deduplication	Combine and clean overlapping data
Versioning	Maintain multiple versions for backtesting
Technical Indicators	Compute SMA, Bollinger Bands, and MACD

This workflow demonstrates how ArcticDB can handle complex financial analysis tasks efficiently. Want more on strategy backtesting or real-time data analysis?