analyse_FTXMOVE.py

#%% 
import glob 
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd 
import seaborn as sns 
sns.set_style('whitegrid')
sns.set_context("paper", 1.5, {"lines.linewidth": 4})

#%% 
# load MOVE data 
flist = glob.glob('data/FTX_OHCLV_2020_BTC-MOVE-0[1-8]*_1min.csv')
# plot 
sumdf = pd.DataFrame()
fig, ax = plt.subplots(figsize=(12,8))
for fname in flist:
    #print(fname)
    df = pd.read_csv(fname)
    fdate = fname.replace('data/FTX_OHCLV_2020_BTC-MOVE-', '').replace('_1min.csv','')
    tmpdf = pd.DataFrame({f'Close_{fdate}':(df.Close.values - df.Close.values[0])},
                            index=((df.TimeStamp - df.TimeStamp.values[0])/60000).astype(int))
    sumdf = pd.concat([sumdf, tmpdf[f'Close_{fdate}']], axis=1)
    tmpdf.reset_index().plot(x='TimeStamp', y=f'Close_{fdate}', 
                            style='-', color='#aaaaaa', lw=1, ax=ax, legend=False)
#display(sumdf)
    

# interporate     
sumdf = sumdf.interpolate()
# calculate median and mean    
sumdf['median'] = sumdf.T.median()
sumdf['mean'] = sumdf.T.mean()
#sumdf['std'] = sumdf.T.std()

sumdf.reset_index().plot(x='index', y=['median', 'mean'], style='-', lw=3, color=['blue', 'orange'], ax=ax)
#sumdf.reset_index().plot(x='index', y=['median', 'mean', 'std'], style='-', lw=3, color=['blue', 'orange', 'green'], ax=ax)
# house keeping 
plt.xlabel('Time [min]')
plt.ylabel('Deviation from close price at t=0')
plt.xlim([0, 2880])
plt.ylim([-500, 500])
plt.savefig('FTX-BTCMOVE-analysisplot1.png')

#%% 
###############
# plot in chronological order 

sumdf = pd.DataFrame()
fig, ax = plt.subplots(figsize=(16,8))
for i, fname in enumerate(flist):
    #fname = flist[0]
    df = pd.read_csv(fname)
    df['tsmin'] = df.TimeStamp/600000
    df['CloseDiff'] = df.Close.values - df.Close.values[0]
    colors = ['tab:red', 'tab:purple', 'tab:brown', 'tab:pink', 'tab:blue', 'tab:orange', 'tab:green']
    df.plot(x='tsmin', y=f'CloseDiff', style='-', color=colors[i%7], lw=1, ax=ax, legend=False)
plt.xlabel('TimeStamp [min]')
plt.ylabel('Deviation from close price at t=0')
plt.savefig('FTX-BTCMOVE-analysisplot2.png')


# %%
# make a histgram of final price 
tmplist = []
for i, fname in enumerate(flist):
    #fname = flist[0]
    df = pd.read_csv(fname)
    df['tsmin'] = df.TimeStamp/600000
    df['CloseDiff'] = df.Close.values - df.Close.values[0]
    colors = ['tab:red', 'tab:purple', 'tab:brown', 'tab:pink', 'tab:blue', 'tab:orange', 'tab:green']
    tmplist += [df.CloseDiff.values[-1]]
tmplist

histdf = pd.DataFrame(tmplist, columns=['Final diff price'])
bins = np.linspace(-200, 200, num=30)
fig, ax = plt.subplots(figsize=(12,8))
sns.distplot(histdf, ax=ax, bins=bins)
plt.title('')
plt.xlim([-200, 200])
plt.xlabel('Final diff price (USD)', fontsize=20)
plt.ylabel('Frequency', fontsize=20)
plt.savefig('FTX-BTCMOVE-histplot.png')