In [59]:

    

import matplotlib
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from os import path
from matplotlib.ticker import ScalarFormatter, FormatStrFormatter, LogFormatterMathtext
#from matplotlib.pyplot import figure, axes, plot, xlabel, ylabel, title, \
#grid, savefig, show


    

In [60]:

    

from tools import FixedOrderFormatter, get_psycinfo_database, GetTemporalPsyc


    

In [61]:

    

%matplotlib inline


    

In [62]:

    

df = pd.DataFrame.from_csv("data/PsycInfo/PsycInfo Articles Review.csv")
df.set_index(df.index.year, inplace=True)


    

In [63]:

    

terms = GetTemporalPsyc()


    

In [64]:

    

len(df), len(terms)


    

    
Out[64]:





(166, 95)

In [65]:

    

terms["Prob"]  = terms.fillna(0).sum(axis=1)
terms["Count"]  = terms["Prob"].values * df.loc[terms.index].values.flatten()


    

In [66]:

    

x = df.index
y = df.values
y_ = np.log2(y)

m1, b1 = np.polyfit(x, y_, 1)
fig = plt.figure(figsize=(9.0, 6.0))
ax = fig.add_subplot(1,1,1)

_ = plt.plot(x, y,  marker='+', label="Publications")
_ = plt.plot(x, 2**(m1*x.values + b1), 'r-', label="Best fit")


z = terms.Count[terms.index >= 1940]
z = z[z.index < 2016]
z_ = np.log2(z.values)
m2, b2 = np.polyfit(z.index, z_, 1)


plt.plot(terms.index, terms.Count, label="Cultural Publications")
_ = plt.plot(z.index, 2**(m2*z.index.values + b2), 'k-', label="Best fit")

ax.set_yscale('log')
ax.yaxis.set_major_formatter(LogFormatterMathtext())
ax.set_ylabel("Articles")
_ = ax.set_xlabel("Year")
plt.tick_params(axis='y', which='minor', labelsize=7)
_ = ax.legend(loc="upper left")


    

    




//anaconda/lib/python3.5/site-packages/matplotlib/scale.py:100: RuntimeWarning: invalid value encountered in less_equal
  a[a <= 0.0] = 1e-300






    

In [67]:

    

1.0/m2, 1.0/m1


    

    
Out[67]:





(6.8449627759337952, array([ 15.49569969]))

In [68]:

    

print (b2)


    

    




-283.061858955

In [69]:

    

print (m2)


    

    




0.146092832457

In [71]:

    

fig = plt.figure(figsize=(9.0, 6.0))
ax = fig.add_subplot(1,1,1)
font = {'family': "Times New Roman",
            'color': 'black',
           'size': 18}
y = df.values / 1000.0
_ = plt.plot(x, y,  marker = 'd', markevery =3, linewidth = 2, label="All PsycINFO Publications", color = 'gray')
y_ = np.log10(y)
m, b = np.polyfit(x, y_, 1)
_ = plt.plot(x, 10**(m*x.values + b), 'r-', linewidth = 2, label="Curve of Best Fit", color = 'black')
ax.set_ylabel("Number of Publications (in thousands)", font)
_ = ax.set_xlabel("Year", font)
plt.tick_params(axis='y', which='minor', labelsize=7)
z = terms.Count[terms.index >= 1940]/1000
z = z[z.index < 2016]
z_ = np.log2(z.values)
plt.plot(terms.index, terms.Count/1000, marker = 'o', markevery = 4, linewidth = 2, label="Psychology Publications with Cultural Terms", color = 'lightslategray')
m2, b2 = np.polyfit(z.index, z_, 1)
#_ = plt.plot(z.index, 2**(m2*z.index.values + b2), 'r-', linewidth = 2, label="Curve of Best Fit", color = 'black')
_ = ax.legend(loc='best')
_ = plt.xlim(1845, 2020)
ax.tick_params(axis='x', labelsize=15)
ax.tick_params(axis='y', labelsize=15)
for tick in ax.get_xticklabels():
    tick.set_fontname("Times New Roman")
for tick in ax.get_yticklabels():
    tick.set_fontname("Times New Roman")
pylab.savefig('PsycINFO1.png')


    

In [15]:

    

y_ = np.log2(y*1000)
m, b = np.polyfit(x, y_, 1)
print("2**({}*x+{})".format(m[0], b[0]))


    

    




2**(0.0645340333298942*x+-112.82389398698189)

In [16]:

    

1.0/m


    

    
Out[16]:





array([ 15.49569969])

In [72]:

    

fig = plt.figure(figsize=(9.0, 6.0))
ax = fig.add_subplot(1,1,1)
font = {'family': "Times New Roman",
            'color': 'black',
           'size': 12}
y = df.values / 1000.0
_ = plt.plot(x, y,  marker = 'd', markevery =4, linewidth = 2, label="All PsycINFO Publications", color = 'gray')
y_ = np.log10(y)
m, b = np.polyfit(x, y_, 1)
_ = plt.plot(x, 10**(m*x.values + b), 'r-', linewidth = 2, label="Curve of Best Fit", color = 'black')
#ax.set_ylabel("Number of Publications (in thousands)", font)
#_ = ax.set_xlabel("Year", font)
plt.tick_params(axis='y', which='major', labelsize=16)
plt.tick_params(axis='x', which='major', labelsize=14)



z = terms.Count[terms.index >= 1940]/1000
z = z[z.index < 2016]
z_ = np.log2(z.values)

plt.plot(terms.index, terms.Count/1000.0, marker = 'o', markevery = 4, linewidth = 2, 
         label="Psychology Publications with Cultural Terms", color = 'lightslategray')

m2, b2 = np.polyfit(z.index, z_, 1)
_ = plt.plot(z.index, 2**(m2*z.index.values + b2), 'r-', linewidth = 2, label="Curve of Best Fit", color = 'black')
#_ = ax.legend(loc='best')
_ = plt.xlim(1845, 2020)
_ = plt.ylim(0, 10)
ax.tick_params(axis='x', labelsize=15)
ax.tick_params(axis='y', labelsize=15)
for tick in ax.get_xticklabels():
    tick.set_fontname("Times New Roman")
for tick in ax.get_yticklabels():
    tick.set_fontname("Times New Roman")
pylab.savefig('PsycINFO2.png')


    

In [ ]: