Kripto paraların Python ile kolerasyonlarını inceleme

from binance import Client
binance_cli = Client(API_KEY, API_SECRET)

tickers = binance_cli.get_exchange_info()

tickers = [a["symbol"] for a in tickers["symbols"]]

usdtPairs = []
btcPairs = []
exclude = ["DOWN", "UP", "BULL", "BEAR", "YEN", "TRY", "EURO", "EUR", "TUSD", "BUSD", "USDC", "USDS"]
for ticker in tickers:
    if ticker.endswith("USDT") and all(element not in ticker for element in exclude):
        usdtPairs.append(ticker)
    elif ticker.endswith("BTC"):
            btcPairs.append(ticker)

usdtPairs

import pandas as pd

"""
    [
        1499040000000,      # Open time
        "0.01634790",       # Open
        "0.80000000",       # High
        "0.01575800",       # Low
        "0.01577100",       # Close
        "148976.11427815",  # Volume
        1499644799999,      # Close time
        "2434.19055334",    # Quote asset volume
        308,                # Number of trades
        "1756.87402397",    # Taker buy base asset volume
        "28.46694368",      # Taker buy quote asset volume
        "17928899.62484339" # Can be ignored
    ]
"""

binance_cli.get_klines(symbol = "BTCUSDT" , interval = '4h', limit  = 10 )

def getData(symbol ,interval = "4h", limit = 10):
    subFrame = pd.DataFrame(binance_cli.get_klines(symbol = symbol , interval = interval, limit  = limit ))
    subFrame = subFrame.iloc[:, :6]
    subFrame.columns = ["Time", "Open", "High", "Low", "Close", "Volume"]
    subFrame = subFrame.set_index("Time")
    subFrame.index = pd.to_datetime(subFrame.index, unit="ms")
    subFrame = subFrame.astype(float)
    return subFrame

arr = []
for pair in usdtPairs[0:50]:
    df = getData(pair)
    print("Getting Klines for", pair)
    arr.append(getData(pair))

dict(zip(usdtPairs[0:50], arr))

df = pd.concat(dict(zip(usdtPairs[0:20], arr)), axis=1)

closedf = df.loc[:, df.columns.get_level_values(1).isin(["Close"])]

closedf.columns = closedf.columns.droplevel(1)

closedf = closedf[closedf.iloc[-1].dropna().index].dropna()

closedf

import numpy as np

"""# Log return formula
    log(percentage change + 1) -> (log is a natural logarithm)
    OR
    log(final element / initial element) -> (log is a natural logarithm)

## Cumulative Return
    exp(sum(all log returns)) - 1
"""

example = pd.Series(np.random.randint(1,5,5))
example.pct_change()
print(np.log(example.pct_change()[2] + 1), np.log(example[2] / example[1]), np.exp(sum(np.log(example.pct_change()[1::] + 1))) - 1)

np.expm1(sum(np.log(example.pct_change()[1::] + 1)))

log_df = np.log(closedf.pct_change() + 1)

log_df.drop(closedf.index[0], inplace=True)

log_df

log_df.corr()

import plotly.express as px

corr = log_df.corr()

fig = px.imshow(corr, color_continuous_scale=px.colors.sequential.Viridis_r)
fig.update_layout(title="Correlation heatmap",
                  height=1000,
                  width=1000,
                  xaxis_nticks=len(corr),
                  yaxis_nticks=len(corr),
                  )
fig.show()