notebook.community

Edit and run



In [19]:

    
import pandas as pd, scipy, matplotlib.pyplot as plt, scipy.optimize as sp, numpy as np
%matplotlib inline



In [329]:

    
c='bitcoin'
df=pd.read_html('http://coinmarketcap.com/currencies/bitcoin/historical-data/?start=20130428&end=20170701')[1]
df['t']=pd.to_datetime(df['Date'])
df=df.sort_index(by='t').reset_index()
df['x']=df.index
df.head(2)









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: FutureWarning: by argument to sort_index is deprecated, pls use .sort_values(by=...)






    Out[329]:







  
    
      
      index
      Date
      Open
      High
      Low
      Close
      Volume
      Market Cap
      t
      x
    
  
  
    
      0
      1524
      Apr 28, 2013
      135.30
      135.98
      132.1
      134.21
      -
      1500520000
      2013-04-28
      0
    
    
      1
      1523
      Apr 29, 2013
      134.44
      147.49
      134.0
      144.54
      -
      1491160000
      2013-04-29
      1



In [330]:

    
def f(x, a, b, c):
    return a*x*x + b*x + c
def g(x, a, b, c):
    return a*np.exp(-b*x+c)
def h(x, a, b):
    return a*(x**b)
def r(x, a, b):
    return a*x+b
def rsq(xdata,ydata,fdata):
    residuals = ydata-fdata
    ss_res = np.sum(residuals**2)
    ss_tot = np.sum((ydata-np.mean(ydata))**2)
    return 1 - (ss_res / ss_tot)



In [331]:

    
fig,ax=plt.subplots(2,2,figsize=(10,7))
z=178
f1=sp.curve_fit(f,df['x'].values,df['Close'].values)
g1=sp.curve_fit(g,df['x'].values,df['Close'].values)
h1=sp.curve_fit(h,df['x'].values,df['Close'].values)
r1=sp.curve_fit(r,df['x'].values,df['Close'].values)
x=np.arange(2,max(df['x'])+z)
x2=np.arange(2,max(df['x'])+z*2)
ax[0][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(df['x'].values,df['Close'].values,f(df['x'].values,*f1[0]))
ax[0][0].plot(x,fdata,lw=3,c='green',label='quadratic, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(df['x'].values,df['Close'].values,g(df['x'].values,*g1[0]))
ax[0][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label='exp, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(df['x'].values,df['Close'].values,h(df['x'].values,*h1[0]))
ax[0][0].plot(x,h(x,*h1[0]),lw=3,c='red',label='power law, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(df['x'].values,df['Close'].values,r(df['x'].values,*r1[0]))
ax[0][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label='linear, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][0].plot(df['x'],df['Close'],c='k',label='BTC all',lw=3)
ax[0][0].legend(loc=2,framealpha=0)

k=365
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[0][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[0][1].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[0][1].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
ax[0][1].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[0][1].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[0][1].legend(loc=2,framealpha=0)

k=365*2
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][0].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
ax[1][0].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][0].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][0].legend(loc=2,framealpha=0)

x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][1].plot(x,fdata,lw=3,c='green')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][1].plot(x,g(x,*g1[0]),lw=3,c='blue')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
ax[1][1].plot(x,h(x,*h1[0]),lw=3,c='red')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][1].plot(x,r(x,*r1[0]),lw=3,c='orange')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][1].legend(loc=2,framealpha=0)
ax[1][1].set_yscale('log')

y2=8000
ax[0][0].set_ylim(100,y2)
ax[0][1].set_ylim(100,y2)
ax[1][0].set_ylim(100,y2)
ax[1][1].set_ylim(100,y2)
plt.savefig(c+'.png')









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in exp
C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in multiply



In [332]:

    
c='ethereum'
df=pd.read_html('http://coinmarketcap.com/currencies/'+c+'/historical-data/?start=20130428&end=20170701')[1]
df['t']=pd.to_datetime(df['Date'])
df=df.sort_index(by='t').reset_index()
df['x']=df.index
df.head(2)









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: FutureWarning: by argument to sort_index is deprecated, pls use .sort_values(by=...)






    Out[332]:







  
    
      
      index
      Date
      Open
      High
      Low
      Close
      Volume
      Market Cap
      t
      x
    
  
  
    
      0
      693
      Aug 07, 2015
      2.83
      3.54
      2.520000
      2.770000
      164329
      -
      2015-08-07
      0
    
    
      1
      692
      Aug 08, 2015
      2.79
      2.80
      0.714725
      0.753325
      674188
      167911000
      2015-08-08
      1



In [333]:

    
fig,ax=plt.subplots(2,2,figsize=(10,7))
z=178
f1=sp.curve_fit(f,df['x'].values,df['Close'].values)
g1=sp.curve_fit(g,df['x'].values,df['Close'].values)
#h1=sp.curve_fit(h,df['x'].values,df['Close'].values)
r1=sp.curve_fit(r,df['x'].values,df['Close'].values)
x=np.arange(2,max(df['x'])+z)
x2=np.arange(2,max(df['x'])+z*2)
ax[0][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(df['x'].values,df['Close'].values,f(df['x'].values,*f1[0]))
ax[0][0].plot(x,fdata,lw=3,c='green',label='quadratic, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(df['x'].values,df['Close'].values,g(df['x'].values,*g1[0]))
ax[0][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label='exp, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(df['x'].values,df['Close'].values,h(df['x'].values,*h1[0]))
#ax[0][0].plot(x,h(x,*h1[0]),lw=3,c='red',label='power law, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
#ax[0][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(df['x'].values,df['Close'].values,r(df['x'].values,*r1[0]))
ax[0][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label='linear, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][0].plot(df['x'],df['Close'],c='k',label=c,lw=3)
ax[0][0].legend(loc=2,framealpha=0)

k=178
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
#h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[0][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[0][1].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[0][1].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[0][1].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
#ax[0][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[0][1].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[0][1].legend(loc=2,framealpha=0)

k=365
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
#h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][0].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[1][0].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
#ax[1][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][0].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][0].legend(loc=2,framealpha=0)

x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][1].plot(x,fdata,lw=3,c='green')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][1].plot(x,g(x,*g1[0]),lw=3,c='blue')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[1][1].plot(x,h(x,*h1[0]),lw=3,c='red')#,label=repr(int(rlabel*10000)/100)+"%")
#ax[1][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][1].plot(x,r(x,*r1[0]),lw=3,c='orange')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][1].legend(loc=2,framealpha=0)
ax[1][1].set_yscale('log')

y1=1
y2=1000
ax[0][0].set_ylim(y1,y2)
ax[0][1].set_ylim(y1,y2)
ax[1][0].set_ylim(y1,y2)
ax[1][1].set_ylim(y1,y2)
plt.savefig(c+'.png')









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in exp
C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in multiply



In [334]:

    
c='litecoin'
df=pd.read_html('http://coinmarketcap.com/currencies/'+c+'/historical-data/?start=20130428&end=20170701')[1]
df['t']=pd.to_datetime(df['Date'])
df=df.sort_index(by='t').reset_index()
df['x']=df.index
df.head(2)









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: FutureWarning: by argument to sort_index is deprecated, pls use .sort_values(by=...)






    Out[334]:







  
    
      
      index
      Date
      Open
      High
      Low
      Close
      Volume
      Market Cap
      t
      x
    
  
  
    
      0
      1524
      Apr 28, 2013
      4.30
      4.40
      4.18
      4.35
      -
      73773400
      2013-04-28
      0
    
    
      1
      1523
      Apr 29, 2013
      4.37
      4.57
      4.23
      4.38
      -
      74952700
      2013-04-29
      1



In [335]:

    
fig,ax=plt.subplots(2,2,figsize=(10,7))
z=178
f1=sp.curve_fit(f,df['x'].values,df['Close'].values)
g1=sp.curve_fit(g,df['x'].values,df['Close'].values)
#h1=sp.curve_fit(h,df['x'].values,df['Close'].values)
r1=sp.curve_fit(r,df['x'].values,df['Close'].values)
x=np.arange(2,max(df['x'])+z)
x2=np.arange(2,max(df['x'])+z*2)
ax[0][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(df['x'].values,df['Close'].values,f(df['x'].values,*f1[0]))
ax[0][0].plot(x,fdata,lw=3,c='green',label='quadratic, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(df['x'].values,df['Close'].values,g(df['x'].values,*g1[0]))
ax[0][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label='exp, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(df['x'].values,df['Close'].values,h(df['x'].values,*h1[0]))
#ax[0][0].plot(x,h(x,*h1[0]),lw=3,c='red',label='power law, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
#ax[0][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(df['x'].values,df['Close'].values,r(df['x'].values,*r1[0]))
ax[0][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label='linear, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][0].plot(df['x'],df['Close'],c='k',label=c,lw=3)
ax[0][0].legend(loc=2,framealpha=0)

k=365
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
#h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[0][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[0][1].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[0][1].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[0][1].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
#ax[0][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[0][1].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[0][1].legend(loc=2,framealpha=0)

k=365*2
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
#h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][0].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[1][0].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
#ax[1][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][0].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][0].legend(loc=2,framealpha=0)

x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][1].plot(x,fdata,lw=3,c='green')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][1].plot(x,g(x,*g1[0]),lw=3,c='blue')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[1][1].plot(x,h(x,*h1[0]),lw=3,c='red')#,label=repr(int(rlabel*10000)/100)+"%")
#ax[1][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][1].plot(x,r(x,*r1[0]),lw=3,c='orange')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][1].legend(loc=2,framealpha=0)
ax[1][1].set_yscale('log')

y1=1
y2=100
ax[0][0].set_ylim(y1,y2)
ax[0][1].set_ylim(y1,y2)
ax[1][0].set_ylim(y1,y2)
ax[1][1].set_ylim(y1,y2)
plt.savefig(c+'.png')









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in exp
C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in multiply



In [336]:

    
c='ripple'
df=pd.read_html('http://coinmarketcap.com/currencies/'+c+'/historical-data/?start=20130428&end=20170701')[1]
df['t']=pd.to_datetime(df['Date'])
df=df.sort_index(by='t').reset_index()
df['x']=df.index
df.head(2)









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: FutureWarning: by argument to sort_index is deprecated, pls use .sort_values(by=...)






    Out[336]:







  
    
      
      index
      Date
      Open
      High
      Low
      Close
      Volume
      Market Cap
      t
      x
    
  
  
    
      0
      1426
      Aug 04, 2013
      0.005874
      0.005927
      0.005874
      0.005882
      -
      45921000
      2013-08-04
      0
    
    
      1
      1425
      Aug 05, 2013
      0.005875
      0.005980
      0.005613
      0.005613
      -
      45928400
      2013-08-05
      1



In [337]:

    
fig,ax=plt.subplots(2,2,figsize=(10,7))
z=178
f1=sp.curve_fit(f,df['x'].values,df['Close'].values)
g1=sp.curve_fit(g,df['x'].values,df['Close'].values)
#h1=sp.curve_fit(h,df['x'].values,df['Close'].values)
r1=sp.curve_fit(r,df['x'].values,df['Close'].values)
x=np.arange(2,max(df['x'])+z)
x2=np.arange(2,max(df['x'])+z*2)
ax[0][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(df['x'].values,df['Close'].values,f(df['x'].values,*f1[0]))
ax[0][0].plot(x,fdata,lw=3,c='green',label='quadratic, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(df['x'].values,df['Close'].values,g(df['x'].values,*g1[0]))
ax[0][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label='exp, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(df['x'].values,df['Close'].values,h(df['x'].values,*h1[0]))
#ax[0][0].plot(x,h(x,*h1[0]),lw=3,c='red',label='power law, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
#ax[0][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(df['x'].values,df['Close'].values,r(df['x'].values,*r1[0]))
ax[0][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label='linear, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][0].plot(df['x'],df['Close'],c='k',label=c,lw=3)
ax[0][0].legend(loc=2,framealpha=0)

k=365/2
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
#h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[0][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[0][1].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[0][1].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[0][1].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
#ax[0][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[0][1].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[0][1].legend(loc=2,framealpha=0)

k=365
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
#h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][0].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[1][0].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
#ax[1][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][0].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][0].legend(loc=2,framealpha=0)

x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][1].plot(x,fdata,lw=3,c='green')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][1].plot(x,g(x,*g1[0]),lw=3,c='blue')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
#ax[1][1].plot(x,h(x,*h1[0]),lw=3,c='red')#,label=repr(int(rlabel*10000)/100)+"%")
#ax[1][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][1].plot(x,r(x,*r1[0]),lw=3,c='orange')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][1].legend(loc=2,framealpha=0)
ax[1][1].set_yscale('log')

y1=0.01
y2=1
ax[0][0].set_ylim(y1,y2)
ax[0][1].set_ylim(y1,y2)
ax[1][0].set_ylim(y1,y2)
ax[1][1].set_ylim(y1,y2)
plt.savefig(c+'.png')









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in exp



In [338]:

    
c='bitcoin'
df1=pd.read_html('http://coinmarketcap.com/currencies/'+c+'/historical-data/?start=20150928&end=20170701')[1]
dt=pd.to_datetime(df1['Date'])
c='ethereum'
df2=pd.read_html('http://coinmarketcap.com/currencies/'+c+'/historical-data/?start=20150928&end=20170701')[1]
c='litecoin'
df3=pd.read_html('http://coinmarketcap.com/currencies/'+c+'/historical-data/?start=20150928&end=20170701')[1]



In [339]:

    
df=df1+df2+df3
df['t']=dt
df=df.sort_index(by='t').reset_index()
df['x']=df.index
df['Close']=df['Market Cap']/1000000.0
df.head(2)
c='market'









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:3: FutureWarning: by argument to sort_index is deprecated, pls use .sort_values(by=...)
  app.launch_new_instance()



In [340]:

    
fig,ax=plt.subplots(2,2,figsize=(10,7))
z=178
f1=sp.curve_fit(f,df['x'].values,df['Close'].values)
g1=sp.curve_fit(g,df['x'].values,df['Close'].values)
h1=sp.curve_fit(h,df['x'].values,df['Close'].values)
r1=sp.curve_fit(r,df['x'].values,df['Close'].values)
x=np.arange(2,max(df['x'])+z)
x2=np.arange(2,max(df['x'])+z*2)
ax[0][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(df['x'].values,df['Close'].values,f(df['x'].values,*f1[0]))
ax[0][0].plot(x,fdata,lw=3,c='green',label='quadratic, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(df['x'].values,df['Close'].values,g(df['x'].values,*g1[0]))
ax[0][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label='exp, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(df['x'].values,df['Close'].values,h(df['x'].values,*h1[0]))
ax[0][0].plot(x,h(x,*h1[0]),lw=3,c='red',label='power law, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(df['x'].values,df['Close'].values,r(df['x'].values,*r1[0]))
ax[0][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label='linear, $R^2=$'+repr(int(rlabel*10000)/100)+"%")
ax[0][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][0].plot(df['x'],df['Close'],c='k',label=c,lw=3)
ax[0][0].legend(loc=2,framealpha=0)

k=int(365*0.75)
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[0][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[0][1].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[0][1].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
ax[0][1].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[0][1].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[0][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[0][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[0][1].legend(loc=2,framealpha=0)

k=int(365*1.5)
f1=sp.curve_fit(f,np.arange(k),df['Close'].values[-k:])
g1=sp.curve_fit(g,np.arange(k),df['Close'].values[-k:])
h1=sp.curve_fit(h,np.arange(k),df['Close'].values[-k:])
r1=sp.curve_fit(r,np.arange(k),df['Close'].values[-k:])
x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][0].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][0].plot(x,fdata,lw=3,c='green',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][0].plot(x,g(x,*g1[0]),lw=3,c='blue',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
ax[1][0].plot(x,h(x,*h1[0]),lw=3,c='red',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][0].plot(x,r(x,*r1[0]),lw=3,c='orange',label=repr(int(rlabel*10000)/100)+"%")
ax[1][0].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][0].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][0].legend(loc=2,framealpha=0)

x=np.arange(2,k+z)
x2=np.arange(2,k+z*2)
ax[1][1].grid()
fdata=f(x,*f1[0])
rlabel=rsq(np.arange(k),df['Close'].values[-k:],f(np.arange(k),*f1[0]))
ax[1][1].plot(x,fdata,lw=3,c='green')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,f(x2,*f1[0]),c='green')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],g(np.arange(k),*g1[0]))
ax[1][1].plot(x,g(x,*g1[0]),lw=3,c='blue')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,g(x2,*g1[0]),c='blue')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],h(np.arange(k),*h1[0]))
ax[1][1].plot(x,h(x,*h1[0]),lw=3,c='red')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,h(x2,*h1[0]),c='red')
rlabel=rsq(np.arange(k),df['Close'].values[-k:],r(np.arange(k),*r1[0]))
ax[1][1].plot(x,r(x,*r1[0]),lw=3,c='orange')#,label=repr(int(rlabel*10000)/100)+"%")
ax[1][1].plot(x2,r(x2,*r1[0]),c='orange')
ax[1][1].plot(np.arange(k),df['Close'][-k:],c='k',label='last '+repr(k),lw=3)
ax[1][1].legend(loc=2,framealpha=0)
ax[1][1].set_yscale('log')

y1=1
y2=200000
ax[0][0].set_ylim(y1,y2)
ax[0][1].set_ylim(y1,y2)
ax[1][0].set_ylim(y1,y2)
ax[1][1].set_ylim(y1,y2)
plt.savefig(c+'.png')









    



C:\Anaconda2\lib\site-packages\ipykernel\__main__.py:4: RuntimeWarning: overflow encountered in exp

	index	Date	Open	High	Low	Close	Volume	Market Cap	t	x
0	1524	Apr 28, 2013	135.30	135.98	132.1	134.21	-	1500520000	2013-04-28	0
1	1523	Apr 29, 2013	134.44	147.49	134.0	144.54	-	1491160000	2013-04-29	1

	index	Date	Open	High	Low	Close	Volume	Market Cap	t	x
0	693	Aug 07, 2015	2.83	3.54	2.520000	2.770000	164329	-	2015-08-07	0
1	692	Aug 08, 2015	2.79	2.80	0.714725	0.753325	674188	167911000	2015-08-08	1

	index	Date	Open	High	Low	Close	Volume	Market Cap	t	x
0	1524	Apr 28, 2013	4.30	4.40	4.18	4.35	-	73773400	2013-04-28	0
1	1523	Apr 29, 2013	4.37	4.57	4.23	4.38	-	74952700	2013-04-29	1

	index	Date	Open	High	Low	Close	Volume	Market Cap	t	x
0	1426	Aug 04, 2013	0.005874	0.005927	0.005874	0.005882	-	45921000	2013-08-04	0
1	1425	Aug 05, 2013	0.005875	0.005980	0.005613	0.005613	-	45928400	2013-08-05	1