MarketStats Engine

A high-performance C++17 engine that processes historical market trade data and computes real-time market statistics in a single pass.

Overview

This project demonstrates low-level C++ optimization and algorithmic efficiency. It reads a CSV containing historical market trades and computes real-time financial metrics for every unique ticker, utilizing a zero-allocation parsing strategy and completely avoiding heavy external dependencies.

Features Computed

For every unique ticker, the engine computes the following metrics incrementally:

• Volume-Weighted Average Price (VWAP)
• High Price
• Low Price
• Total Volume
• Total Traded Value
• Number of Trades

Architecture & Complexity

The core engine utilizes a hash-based aggregation strategy (std::unordered_map<std::string, MarketMetrics>). As new trades are ingested, metrics are updated incrementally in constant time.

• Time Complexity: O(N), where N is the total number of trades (rows) processed.
• Memory Complexity: O(K), where K is the number of unique tickers. This ensures the memory footprint remains minimal and strictly bounded, regardless of total trade volume.

C++17 Optimizations

To achieve maximum throughput, standard standard-library parsing bottlenecks were eliminated:

• Zero-Allocation String Parsing: Utilizes std::string_view to extract ticker and numerical data directly from the CSV read buffer, entirely bypassing std::string heap allocations and copies.
• High-Speed Numeric Conversion: Replaces std::stringstream and std::stod with C++17's std::from_chars, providing low-level, locale-independent, and zero-allocation float and integer parsing.

Benchmark Results

The engine was profiled using dynamically generated mock market data on a single consumer-grade CPU thread.

Rows Processed	Execution Time	Throughput (Rows / Second)
100,000	0.02 seconds	~4.91 Million
500,000	0.06 seconds	~7.98 Million
1,000,000	0.11 seconds	~8.89 Million

Result: Processing 1,000,000 rows of market data executes in exactly 110 milliseconds, achieving nearly 8.9 million rows per second on a single thread.

Build Instructions

mkdir build
cd build
cmake ..
cmake --build .
cd ..
.\build\MarketStatsEngine.exe

Sample Output

Ticker    VWAP        High        Low         Volume      Trades
-----------------------------------------------------------------
AAPL      168.24      196.99      115.60      4000        6
AMZN      148.54      170.81      113.95      2600        5
GOOG      142.43      195.07      109.77      900         4
MSFT      134.83      178.52      104.64      2200        3
TSLA      168.92      180.84      121.23      500         2