What is algorithmic Trading ?!
In simple words, algorithmic trading is the use of programs to make systematic trades based on the strategy implemented by the programming language.
algorithmic trading can be fully-automated, semi-automated, or give signals to be executed manually.
In this article, we will develop a simple semi-automated strategy using Binance API to get data and execute trades.
I am working with Jupiter notebook, using python, pandas, and NumPy
Starting with importing the libraries needed install python-binance to connect to Binance API.
I am using binance testnet API to make anyone able to go and execute the code even if you don’t have binance account.
first, you have to have an API KEY and SECRET KEY to provide to the client function to establish a connection
to generate an API key and secret key go to Binance Spot Test Network sign with GitHub and generate HMAC SHA256 key.
finally, replace APIKEY and SECRETKEY variables in the code with actual keys.
import pandas as pd import numpy as np !pip install python-binance from binance import Client APIKEY = 'Your API KEY' SECRETKEY = 'Your Secret KEY' client =Client(APIKEY,SECRETKEY) client.API_URL = 'https://testnet.binance.vision/api' acc = client.get_account() acc after establishing connection let’s see our TEST account balance Info with client.get_account() function, and here is the result.
--- {'makerCommission': 0, 'takerCommission': 0, 'buyerCommission': 0, 'sellerCommission': 0, 'canTrade': True, 'canWithdraw': False, 'canDeposit': False, 'updateTime': 1644598822834, 'accountType': 'SPOT', 'balances': [{'asset': 'BNB', 'free': '1000.00000000', 'locked': '0.00000000'}, {'asset': 'BTC', 'free': '1.01300000', 'locked': '0.00000000'}, {'asset': 'BUSD', 'free': '10000.00000000', 'locked': '0.00000000'}, {'asset': 'ETH', 'free': '100.00000000', 'locked': '0.00000000'}, {'asset': 'LTC', 'free': '500.00000000', 'locked': '0.00000000'}, {'asset': 'TRX', 'free': '500000.00000000', 'locked': '0.00000000'}, {'asset': 'USDT', 'free': '9449.55739000', 'locked': '0.00000000'}, {'asset': 'XRP', 'free': '50000.00000000', 'locked': '0.00000000'}], 'permissions': ['SPOT']} Now it’s time to get crypto data stream
we have more than one option about how to get data from Binance, in this article I want to make it simple enough so I am using
client.get_historical_klines() function
Parameters provided to this function are:
symbol: which is the cryptocurrency symbol you want to trade in our case its BTCUSDT
Time Frame: time frame which you’ll trade in our case I chose a 1minute time frame.
Period: How many past periods to get in our case I chose 20 minutes.
def dataorg(symbol): data = pd.DataFrame(client.get_historical_klines( symbol , '1m', '20 min ago UTC')) df = data.iloc[:,:6] df.columns = ['Time', 'Open', 'High' , 'Low', 'Close','Volume'] df = df.set_index('Time') df.index = pd.to_datetime(df.index,unit='ms') df = df.astype(float) return df dataorg('BTCUSDT') I defined a function that takes a symbol argument to organize data,
and make a data frame out of it
here is the resulting data frame containing:
Time , Open Price , Low Price , Close Price and Volume.
Here we want to calculate the cumulative return of the last 20 minutes to start work on the trading strategy.
Cumulative return formula: (1+ 1min return)*(1+cumlative return of last min)-1
pct_change(): calcluate the change in price in the last period.
cumprod(): calculate the second half of the formula.
test = dataorg('BTCUSDT') ret = (test.Open.pct_change() + 1).cumprod() -1 ret Time 2022-02-11 16:52:00 NaN 2022-02-11 16:53:00 0.000279 2022-02-11 16:54:00 0.000974 2022-02-11 16:55:00 0.001551 2022-02-11 16:56:00 0.001645 2022-02-11 16:57:00 0.002303 2022-02-11 16:58:00 0.002782 2022-02-11 16:59:00 0.002060 2022-02-11 17:00:00 0.001818 2022-02-11 17:01:00 0.000858 2022-02-11 17:02:00 0.001216 2022-02-11 17:03:00 0.002534 2022-02-11 17:04:00 0.002666 2022-02-11 17:05:00 0.003242 2022-02-11 17:06:00 0.002548 2022-02-11 17:07:00 0.001416 2022-02-11 17:08:00 0.002460 2022-02-11 17:09:00 0.002109 2022-02-11 17:10:00 0.002550 2022-02-11 17:11:00 0.003376 Name: Open, dtype: float64 Strategy implementation
We will develop a trend following strategy which as simple as:
If the price of Bitcoin increases by more than a certain percentage (0.2%) then BUY order and after buy If price increased or decreased by certain percentage ( 0.2% , -0.2%) then SELL.
def strategy(symbol,qty,inPosition=False): while True: df = dataorg(symbol) cumret = (df.Open.pct_change()+1).cumprod()-1 if not inPosition: if cumret[-1] > 0.002: ordertime = df.index[-1] order = client.create_order(symbol=symbol, side='BUY',type='MARKET', quantity=qty) print(order) inPosition=True break else: print('No trade excuted') if inPosition: while True: df=dataorg(symbol) afterbuy = df.loc[df.index > pd.to_datetime( order['transactTime'], unit='ms')] if len(afterbuy) > 0: afterbuyret = (df.Open.pct_change()+1).cumprod()-1 print(afterbuyret) if afterbuyret[-1]>0.002 or afterbuyret[-1]< -0.002: sellorder = client.create_order(symbol=symbol, side='SELL',type='MARKET', quantity=qty) print(sellorder) break it’s time to see the prices in the last 20 mins in a plot.
test.Open.plot() 
It seems like we have an uptrend .
Trade execution
Our final step is to call strategy() function and wait for a trade to be filled,
In this case we have a buy order filled then list of cumulative return after a trade and when the return exceed our threshold (+0.2% or -0.2%) then a sell trade been filled
and it’s a profitable trade 😀
strategy('BTCUSDT', 0.001) {'symbol': 'BTCUSDT', 'orderId': 4982968, 'orderListId': -1, 'clientOrderId': 'oappYtNU1466jYiXCV3URd', 'transactTime': 1644599469117, 'price': '0.00000000', 'origQty': '0.00100000', 'executedQty': '0.00100000', 'cummulativeQuoteQty': '43.65218000', 'status': 'FILLED', 'timeInForce': 'GTC', 'type': 'MARKET', 'side': 'BUY', 'fills': [{'price': '43652.18000000', 'qty': '0.00100000', 'commission': '0.00000000', 'commissionAsset': 'BTC', 'tradeId': 1240296}]} Time 2022-02-11 16:52:00 NaN 2022-02-11 16:53:00 0.000279 2022-02-11 16:54:00 0.000974 2022-02-11 16:55:00 0.001551 2022-02-11 16:56:00 0.001645 2022-02-11 16:57:00 0.002303 2022-02-11 16:58:00 0.002782 2022-02-11 16:59:00 0.002060 2022-02-11 17:00:00 0.001818 2022-02-11 17:01:00 0.000858 2022-02-11 17:02:00 0.001216 2022-02-11 17:03:00 0.002534 2022-02-11 17:04:00 0.002666 2022-02-11 17:05:00 0.003242 2022-02-11 17:06:00 0.002548 2022-02-11 17:07:00 0.001416 2022-02-11 17:08:00 0.002460 2022-02-11 17:09:00 0.002109 2022-02-11 17:10:00 0.002550 2022-02-11 17:11:00 0.003376 2022-02-11 17:12:00 0.003520 Name: Open, dtype: float64 {'symbol': 'BTCUSDT', 'orderId': 4983220, 'orderListId': -1, 'clientOrderId': 'OWLlWGmI044UlKEFaURffy', 'transactTime': 1644599521374, 'price': '0.00000000', 'origQty': '0.00100000', 'executedQty': '0.00100000', 'cummulativeQuoteQty': '43.65542000', 'status': 'FILLED', 'timeInForce': 'GTC', 'type': 'MARKET', 'side': 'SELL', 'fills': [{'price': '43655.42000000', 'qty': '0.00100000', 'commission': '0.00000000', 'commissionAsset': 'USDT', 'tradeId': 1240505}]} what's next
Ok guys, that's it for this article of course this isn't the best trading bot to make profits and it lacks a lot,
especially data anlysis and backtesting ,
I'll provide more insights in my upcoming article,
I hope you enjoy.
Top comments (1)
Some comments may only be visible to logged-in visitors. Sign in to view all comments.