AI

Implement Dollar Bars Sampling

Medium
Reinforcement Learning

Implement a function that generates dollar bars from trade data. Dollar bars are an alternative sampling method used in financial machine learning where a new bar (candlestick) is created whenever a specified dollar amount has been traded, rather than at fixed time intervals.

Given a list of trades where each trade is represented as a tuple (price, volume), and a dollar threshold, your function should:

  1. Calculate the dollar value of each trade as price * volume
  2. Accumulate dollar values until the threshold is reached or exceeded
  3. When the threshold is reached, create a bar containing:
    • open: The price of the first trade in the bar
    • high: The highest price in the bar
    • low: The lowest price in the bar
    • close: The price of the last trade in the bar
    • volume: Total volume traded in the bar
    • dollar_value: Total dollar value traded in the bar
  4. Reset the accumulator and continue with the remaining trades

Return a list of tuples, where each tuple represents a bar in the format (open, high, low, close, volume, dollar_value). All values should be rounded to 4 decimal places.

Note: Trades that do not accumulate enough dollar value to meet the threshold are not included in any bar.

Examples

Example 1:
Input: trades = [(100.0, 10), (101.0, 15), (99.0, 20), (102.0, 5)], dollar_threshold = 2000.0
Output: [(100.0, 101.0, 100.0, 101.0, 25, 2515.0), (99.0, 102.0, 99.0, 102.0, 25, 2490.0)]
Explanation: Processing trades sequentially: 1. Trade 1: price=100.0, volume=10 -> dollar=1000.0, accumulated=1000.0 (below threshold) 2. Trade 2: price=101.0, volume=15 -> dollar=1515.0, accumulated=2515.0 (>= 2000, create bar) - Bar 1: open=100.0, high=101.0, low=100.0, close=101.0, volume=25, dollar=2515.0 - Reset accumulator 3. Trade 3: price=99.0, volume=20 -> dollar=1980.0, accumulated=1980.0 (below threshold) 4. Trade 4: price=102.0, volume=5 -> dollar=510.0, accumulated=2490.0 (>= 2000, create bar) - Bar 2: open=99.0, high=102.0, low=99.0, close=102.0, volume=25, dollar=2490.0 Result: Two bars capturing the market activity based on dollar volume traded.

Starter Code

def dollar_bars(trades: list, dollar_threshold: float) -> list:
    """
    Generate dollar bars from trade data.
    
    Args:
        trades: List of tuples (price, volume) representing individual trades
        dollar_threshold: Dollar amount threshold for creating a new bar
        
    Returns:
        List of dollar bars as tuples (open, high, low, close, volume, dollar_value)
    """
    # Your code here
    pass
Lines: 1Characters: 0
Ready
The AI Interview - Master AI/ML Interviews