Technical-Indicators

Overview

1. A python script gets data for a stock using the yfinance library with specified period and intervals. Data is saved into a CSV file.

2. Then, readCSV() in csv.cpp reads the file and formats the data.

3. A list of indicators like Bollinger Bands, SMA (Simple Moving Average), VWAP (Volume Weighted Average Price), and EMA (Exponential Moving Average) are defined to run on the stock's OHLC and Volume data.

4. Use cmake command to compile and then run.

Speed Test

20-SMA, 100-SMA, and 250-SMA on Apple's 10 year data with a single data as interval.

get_data.py ->

df = download_data("AAPL", "10y", "1d")

st = time()
moving_avg_20 = df['Close'].rolling(window=20).mean()
moving_avg_100 = df['Close'].rolling(window=100).mean()
moving_avg_250 = df['Close'].rolling(window=250).mean()
et = time()
print(f"Execution time : {et-st} seconds")

Technical-Indicators % python3 get_data.py
Got data for AAPL
Price           Close       High        Low       Open     Volume
Date                                                             
2016-02-04  21.898483  22.063969  21.578848  21.730731  185886800
2016-02-05  21.313620  21.971028  21.238812  21.880351  185672400
2016-02-08  21.538048  21.694464  21.091463  21.111865  216085600
2016-02-09  21.533508  21.748867  21.293215  21.374824  177324800
2016-02-10  21.370287  21.841807  21.331750  21.744330  169374400
Downloaded
Execution time : 0.002521991729736328 seconds

main.cpp ->

int main() {
    // ... Reading CSV file

    // 20-SMA
    xarray<double> sma_20 = SimpleMovingAverage(20, csvData.Close);
    std::cout << "20-SMA vector size: " << sma_20.size() << std::endl;
    // 100-SMA
    xarray<double> sma_100 = SimpleMovingAverage(100, csvData.Close);
    std::cout << "100-SMA vector size: " << sma_100.size() << std::endl;
    // 250-SMA
    xarray<double> sma_250 = SimpleMovingAverage(250, csvData.Close);
    std::cout << "250-SMA vector size: " << sma_250.size() << std::endl;
    
    // ... Rest of the code
}

Technical-Indicators % cmake --build ./build --config Release && ./build/main
[ 50%] Building CXX object CMakeFiles/main.dir/main.cpp.o
[100%] Linking CXX executable main
[100%] Built target main
Headers: Date,Close,High,Low,Open,Volume
Total data points: 2514
20-SMA vector size: 2495
100-SMA vector size: 2415
250-SMA vector size: 2265
Execution time: 0.099959 ms

So, we can see here for the same query Python code using Numpy at the backend to calculate the MA is fast, only 2.5 ms. But, the xtensor framework in C++ is just faster, taking 0.099 ms.

Installation Steps

1. I have used the xtensor library (it's like Numpy but for C++) to handle arrays and calculations. To install xtensor -

   git clone https://github.com/microsoft/vcpkg.git
   
   ./vcpkg/bootstrap-vcpkg.sh
   
   ./vcpkg/vcpkg install xtensor

2. Inside this project folder on your terminal, create a 'build' folder and move in it. mkdir build && cd build

3. Configure the project

   1. Debug mode ( Takes longer to run as compared to Release mode)

      cmake .. -DCMAKE_TOOLCHAIN_FILE=/path/to/vcpkg/scripts/buildsystems/vcpkg.cmake && cd ..

   2. Release mode

      cmake .. -DCMAKE_BUILD_TYPE=Release \           
          -DCMAKE_TOOLCHAIN_FILE=/path/to/vcpkg/scripts/buildsystems/vcpkg.cmake && cd ..

4. Compile and Run

   1. Debug mode

      cmake --build ./build && ./build/main

   2. Release mode

      cmake --build ./build --config Release && ./build/main

5. Removing the 'build' folder rm -rf build

6. To get the financial data, we use yfinance and for plotting we use matplotlib.

   pip3 install yfinance matplotlib

Pointers

1. Add more indicators like RSI, Average True Range, and MACD.

2. Find a way to use these functions (SMA, EMA, and others) through Python directly in code.

3. Find Buy, Hold or Sell signals using these indicators and stock price.