Fastquant - Backtest and optimize your trading strategies with only 3 lines of code!

Overview

fastquant šŸ¤“

Build Status Code style: black License: MIT Downloads

Bringing backtesting to the mainstream

fastquant allows you to easily backtest investment strategies with as few as 3 lines of python code. Its goal is to promote data driven investments by making quantitative analysis in finance accessible to everyone.

To do this type of analysis without coding, you can also try out Hawksight, which was just recently launched! šŸ˜„

If you want to interact with us directly, you can also reach us on the Hawksight discord. Feel free to ask about fastquant in the #feedback-suggestions and #bug-report channels.

Features

  1. Easily access historical stock data
  2. Backtest and optimize trading strategies with only 3 lines of code

* - Both Yahoo Finance and Philippine stock data data are accessible straight from fastquant

Check out our blog posts in the fastquant website and this intro article on Medium!

Installation

Python

pip install fastquant
or
python -m pip install fastquant

Get stock data

All symbols from Yahoo Finance and Philippine Stock Exchange (PSE) are accessible via get_stock_data.

Python

from fastquant import get_stock_data
df = get_stock_data("JFC", "2018-01-01", "2019-01-01")
print(df.head())

#           dt  close
#   2019-01-01  293.0
#   2019-01-02  292.0
#   2019-01-03  309.0
#   2019-01-06  323.0
#   2019-01-07  321.0

Get crypto data

The data is pulled from Binance, and all the available tickers are found here.

Python

from fastquant import get_crypto_data
crypto = get_crypto_data("BTC/USDT", "2018-12-01", "2019-12-31")
crypto.head()

#             open    high     low     close    volume
# dt                                                          
# 2018-12-01  4041.27  4299.99  3963.01  4190.02  44840.073481
# 2018-12-02  4190.98  4312.99  4103.04  4161.01  38912.154790
# 2018-12-03  4160.55  4179.00  3827.00  3884.01  49094.369163
# 2018-12-04  3884.76  4085.00  3781.00  3951.64  48489.551613
# 2018-12-05  3950.98  3970.00  3745.00  3769.84  44004.799448

Backtest trading strategies

Simple Moving Average Crossover (15 day MA vs 40 day MA)

Daily Jollibee prices from 2018-01-01 to 2019-01-01

from fastquant import backtest
backtest('smac', df, fast_period=15, slow_period=40)

# Starting Portfolio Value: 100000.00
# Final Portfolio Value: 102272.90

Want to do this without coding at all?

If you want to make this kind of analysis even more simple without having to code at all (or want to avoid the pain of doing all of the setup required), you can signup for free and try out Hawksight - this new no-code tool I’m building to democratize data driven investments.

Hoping to make these kinds of powerful analyses accessible to more people!

Optimize trading strategies with automated grid search

fastquant allows you to automatically measure the performance of your trading strategy on multiple combinations of parameters. All you need to do is to input the values as iterators (like as a list or range).

Simple Moving Average Crossover (15 to 30 day MA vs 40 to 55 day MA)

Daily Jollibee prices from 2018-01-01 to 2019-01-01

from fastquant import backtest
res = backtest("smac", df, fast_period=range(15, 30, 3), slow_period=range(40, 55, 3), verbose=False)

# Optimal parameters: {'init_cash': 100000, 'buy_prop': 1, 'sell_prop': 1, 'execution_type': 'close', 'fast_period': 15, 'slow_period': 40}
# Optimal metrics: {'rtot': 0.022, 'ravg': 9.25e-05, 'rnorm': 0.024, 'rnorm100': 2.36, 'sharperatio': None, 'pnl': 2272.9, 'final_value': 102272.90}

print(res[['fast_period', 'slow_period', 'final_value']].head())

#	fast_period	slow_period	final_value
#0	15	        40	        102272.90
#1	21	        40	         98847.00
#2	21	        52	         98796.09
#3	24	        46	         98008.79
#4	15	        46	         97452.92

Library of trading strategies

Strategy Alias Parameters
Relative Strength Index (RSI) rsi rsi_period, rsi_upper, rsi_lower
Simple moving average crossover (SMAC) smac fast_period, slow_period
Exponential moving average crossover (EMAC) emac fast_period, slow_period
Moving Average Convergence Divergence (MACD) macd fast_perod, slow_upper, signal_period, sma_period, dir_period
Bollinger Bands bbands period, devfactor
Buy and Hold buynhold N/A
Sentiment Strategy sentiment keyword , page_nums, senti
Custom Prediction Strategy custom upper_limit, lower_limit, custom_column
Custom Ternary Strategy ternary buy_int, sell_int, custom_column

Relative Strength Index (RSI) Strategy

backtest('rsi', df, rsi_period=14, rsi_upper=70, rsi_lower=30)

# Starting Portfolio Value: 100000.00
# Final Portfolio Value: 132967.87

Simple moving average crossover (SMAC) Strategy

backtest('smac', df, fast_period=10, slow_period=30)

# Starting Portfolio Value: 100000.00
# Final Portfolio Value: 95902.74

Exponential moving average crossover (EMAC) Strategy

backtest('emac', df, fast_period=10, slow_period=30)

# Starting Portfolio Value: 100000.00
# Final Portfolio Value: 90976.00

Moving Average Convergence Divergence (MACD) Strategy

backtest('macd', df, fast_period=12, slow_period=26, signal_period=9, sma_period=30, dir_period=10)

# Starting Portfolio Value: 100000.00
# Final Portfolio Value: 96229.58

Bollinger Bands Strategy

backtest('bbands', df, period=20, devfactor=2.0)

# Starting Portfolio Value: 100000.00
# Final Portfolio Value: 97060.30

News Sentiment Strategy

Use Tesla (TSLA) stock from yahoo finance and news articles from Business Times

from fastquant import get_yahoo_data, get_bt_news_sentiment
data = get_yahoo_data("TSLA", "2020-01-01", "2020-07-04")
sentiments = get_bt_news_sentiment(keyword="tesla", page_nums=3)
backtest("sentiment", data, sentiments=sentiments, senti=0.2)

# Starting Portfolio Value: 100000.00
# Final Portfolio Value: 313198.37
# Note: Unfortunately, you can't recreate this scenario due to inconsistencies in the dates and sentiments that is scraped by get_bt_news_sentiment. In order to have a quickstart with News Sentiment Strategy you need to make the dates consistent with the sentiments that you are scraping.

from fastquant import get_yahoo_data, get_bt_news_sentiment
from datetime import datetime, timedelta

# we get the current date and delta time of 30 days
current_date = datetime.now().strftime("%Y-%m-%d")
delta_date = (datetime.now() - timedelta(30)).strftime("%Y-%m-%d")
data = get_yahoo_data("TSLA", delta_date, current_date)
sentiments = get_bt_news_sentiment(keyword="tesla", page_nums=3)
backtest("sentiment", data, sentiments=sentiments, senti=0.2)

Multi Strategy

Multiple registered strategies can be utilized together in an OR fashion, where buy or sell signals are applied when at least one of the strategies trigger them.

df = get_stock_data("JFC", "2018-01-01", "2019-01-01")

# Utilize single set of parameters
strats = { 
    "smac": {"fast_period": 35, "slow_period": 50}, 
    "rsi": {"rsi_lower": 30, "rsi_upper": 70} 
} 
res = backtest("multi", df, strats=strats)
res.shape
# (1, 16)


# Utilize auto grid search
strats_opt = { 
    "smac": {"fast_period": 35, "slow_period": [40, 50]}, 
    "rsi": {"rsi_lower": [15, 30], "rsi_upper": 70} 
} 

res_opt = backtest("multi", df, strats=strats_opt)
res_opt.shape
# (4, 16)

Custom Strategy for Backtesting Machine Learning & Statistics Based Predictions

This powerful strategy allows you to backtest your own trading strategies using any type of model w/ as few as 3 lines of code after the forecast!

Predictions based on any model can be used as a custom indicator to be backtested using fastquant. You just need to add a custom column in the input dataframe, and set values for upper_limit and lower_limit.

The strategy is structured similar to RSIStrategy where you can set an upper_limit, above which the asset is sold (considered "overbought"), and a lower_limit, below which the asset is bought (considered "underbought). upper_limit is set to 95 by default, while lower_limit is set to 5 by default.

In the example below, we show how to use the custom strategy to backtest a custom indicator based on out-of-sample time series forecasts. The forecasts were generated using Facebook's Prophet package on Bitcoin prices.

from fastquant import get_crypto_data, backtest
from fbprophet import Prophet
import pandas as pd
from matplotlib import pyplot as plt

# Pull crypto data
df = get_crypto_data("BTC/USDT", "2019-01-01", "2020-05-31")

# Fit model on closing prices
ts = df.reset_index()[["dt", "close"]]
ts.columns = ['ds', 'y']
m = Prophet(daily_seasonality=True, yearly_seasonality=True).fit(ts)
forecast = m.make_future_dataframe(periods=0, freq='D')

# Predict and plot
pred = m.predict(forecast)
fig1 = m.plot(pred)
plt.title('BTC/USDT: Forecasted Daily Closing Price', fontsize=25)

+1.5%, and sell when it's < -1.5%. df["custom"] = expected_1day_return.multiply(-1) backtest("custom", df.dropna(),upper_limit=1.5, lower_limit=-1.5)">
# Convert predictions to expected 1 day returns
expected_1day_return = pred.set_index("ds").yhat.pct_change().shift(-1).multiply(100)

# Backtest the predictions, given that we buy bitcoin when the predicted next day return is > +1.5%, and sell when it's < -1.5%.
df["custom"] = expected_1day_return.multiply(-1)
backtest("custom", df.dropna(),upper_limit=1.5, lower_limit=-1.5)

See more examples here.

fastquant API

View full list of fastquan API here

Be part of the growing fastquant community

Want to discuss more about fastquant with other users, and our team of developers?

You can reach us on the Hawksight discord. Feel free to ask about fastquant in the #feedback-suggestions and #bug-report channels.

Run fastquant in a Docker Container

>> df.head()">
# Build the image
docker build -t myimage .

# Run the container
docker run -t -d -p 5000:5000 myimage

# Get the container id
docker ps

# SSH into the fastquant container
docker exec -it 
   
     /bin/bash

# Run python and use fastquant
python

>>> from fastquant import get_stock_data
>>> df = get_stock_data("TSLA", "2019-01-01", "2020-01-01")
>>> df.head()

   
Owner
Lorenzo Ampil
co-founder & dev @ Hawksight.co | democratizing smart defi | creator of fastquant | top contributor @flipsidecrypto | šŸ‡µšŸ‡­ based in šŸ‡øšŸ‡¬
Lorenzo Ampil
[AAAI22] Reliable Propagation-Correction Modulation for Video Object Segmentation

Reliable Propagation-Correction Modulation for Video Object Segmentation (AAAI22) Preview version paper of this work is available at: https://arxiv.or

Xiaohao Xu 70 Dec 04, 2022
An executor that performs image segmentation on fashion items

ClothingSegmenter U2NET fashion image/clothing segmenter based on https://github.com/levindabhi/cloth-segmentation Overview The ClothingSegmenter exec

Jina AI 5 Mar 30, 2022
Implementation of "Deep Implicit Templates for 3D Shape Representation"

Deep Implicit Templates for 3D Shape Representation Zerong Zheng, Tao Yu, Qionghai Dai, Yebin Liu. arXiv 2020. This repository is an implementation fo

Zerong Zheng 144 Dec 07, 2022
This is an official pytorch implementation of Fast Fourier Convolution.

Fast Fourier Convolution (FFC) for Image Classification This is the official code of Fast Fourier Convolution for image classification on ImageNet. Ma

pkumi 199 Jan 03, 2023
Implementation of ResMLP, an all MLP solution to image classification, in Pytorch

ResMLP - Pytorch Implementation of ResMLP, an all MLP solution to image classification out of Facebook AI, in Pytorch Install $ pip install res-mlp-py

Phil Wang 178 Dec 02, 2022
Accelerating BERT Inference for Sequence Labeling via Early-Exit

Sequence-Labeling-Early-Exit Code for ACL 2021 paper: Accelerating BERT Inference for Sequence Labeling via Early-Exit Requirement: Please refer to re

ęŽå­ē”· 23 Oct 14, 2022
Automatic voice-synthetised summaries of latest research papers on arXiv

PaperWhisperer PaperWhisperer is a Python application that keeps you up-to-date with research papers. How? It retrieves the latest articles from arXiv

Valerio Velardo 124 Dec 20, 2022
Cascaded Pyramid Network (CPN) based on Keras (Tensorflow backend)

ML2 Takehome Project Reimplementing the paper: Cascaded Pyramid Network for Multi-Person Pose Estimation Dataset The model uses the COCO dataset which

Vo Van Tu 1 Nov 22, 2021
The official repo of the CVPR 2021 paper Group Collaborative Learning for Co-Salient Object Detection .

GCoNet The official repo of the CVPR 2021 paper Group Collaborative Learning for Co-Salient Object Detection . Trained model Download final_gconet.pth

Qi Fan 46 Nov 17, 2022
Python library containing BART query generation and BERT-based Siamese models for neural retrieval.

Neural Retrieval Embedding-based Zero-shot Retrieval through Query Generation leverages query synthesis over large corpuses of unlabeled text (such as

Amazon Web Services - Labs 35 Apr 14, 2022
Memory Defense: More Robust Classificationvia a Memory-Masking Autoencoder

Memory Defense: More Robust Classificationvia a Memory-Masking Autoencoder Authors: - Eashan Adhikarla - Dan Luo - Dr. Brian D. Davison Abstract Many

Eashan Adhikarla 4 Dec 25, 2022
Make your AirPlay devices as TTS speakers

Apple AirPlayer Home Assistant integration component, make your AirPlay devices as TTS speakers. Before Use 2021.6.X or earlier Apple Airplayer compon

George Zhao 117 Dec 15, 2022
Pynomial - a lightweight python library for implementing the many confidence intervals for the risk parameter of a binomial model

Pynomial - a lightweight python library for implementing the many confidence intervals for the risk parameter of a binomial model

Demetri Pananos 9 Oct 04, 2022
Torch implementation of various types of GAN (e.g. DCGAN, ALI, Context-encoder, DiscoGAN, CycleGAN, EBGAN, LSGAN)

gans-collection.torch Torch implementation of various types of GANs (e.g. DCGAN, ALI, Context-encoder, DiscoGAN, CycleGAN, EBGAN). Note that EBGAN and

Minchul Shin 53 Jan 22, 2022
[ACMMM 2021, Oral] Code release for "Elastic Tactile Simulation Towards Tactile-Visual Perception"

EIP: Elastic Interaction of Particles Code release for "Elastic Tactile Simulation Towards Tactile-Visual Perception", in ACMMM (Oral) 2021. By Yikai

Yikai Wang 37 Dec 20, 2022
Nested Graph Neural Network (NGNN) is a general framework to improve a base GNN's expressive power and performance

Nested Graph Neural Networks About Nested Graph Neural Network (NGNN) is a general framework to improve a base GNN's expressive power and performance.

Muhan Zhang 38 Jan 05, 2023
Offline Multi-Agent Reinforcement Learning Implementations: Solving Overcooked Game with Data-Driven Method

Overcooked-AI We suppose to apply traditional offline reinforcement learning technique to multi-agent algorithm. In this repository, we implemented be

Baek In-Chang 14 Sep 16, 2022
Here is the diagnostic tool for BMVC 2021 paper Diagnosing Errors in Video Relation Detectors.

Here is the diagnostic tool for BMVC 2021 paper Diagnosing Errors in Video Relation Detectors. We provide a tiny ground truth file demo_gt.json, and t

Shuo Chen 3 Dec 26, 2022
The official repository for Deep Image Matting with Flexible Guidance Input

FGI-Matting The official repository for Deep Image Matting with Flexible Guidance Input. Paper: https://arxiv.org/abs/2110.10898 Requirements easydict

Hang Cheng 51 Nov 10, 2022
The world's simplest facial recognition api for Python and the command line

Face Recognition You can also read a translated version of this file in Chinese ē®€ä½“äø­ę–‡ē‰ˆ or in Korean ķ•œźµ­ģ–“ or in Japanese ę—„ęœ¬čŖž. Recognize and manipulate fa

Adam Geitgey 46.9k Jan 03, 2023