import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

Grab some data on transistor size/count

https://en.wikipedia.org/wiki/Transistor_count

from wikitables import import_tables
tables = import_tables('Transistor_count')
print(tables[0])

<WikiTable 'Transistor count[0]'>

Parse it (ignore this if you wish)

This grabs the data from the Wikipedia article.

def getfloat(value):
    try:
        return float(value)
    except ValueError:
        return 0.0

table = tables[0]

years = [r['Date of introduction'].value for r in table.rows]
process = [r['Process'].value for r in table.rows]
process = [getfloat(p.split()[0].replace(',','')) if p != '' else 0 for p in process]
count = [r['Transistor count'].value for r in table.rows]
count = [c if type(c) is int else 0.0 if len(c.replace(',','').split()) == 0 else [float(di) for di in c.replace(',','').split() if di.isdigit()][0] for c in count]


years = np.array(years)
process = np.array(process)
count = np.array(count)

I = np.where((process!=0) & (count!=0))
years = years[I]
process = process[I]
count = count[I]

Let's plot it!

Is this the best way to display the data?

What would you change?

plt.figure(figsize=(10,5))
plt.plot(years, process, 'o')
plt.xlabel('year')
plt.ylabel('nm')

Text(0,0.5,'nm')

plt.figure(figsize=(10,5))
plt.plot(years, count, 'o') #, basey=2)
plt.xlabel('year')
plt.ylabel('count')

Text(0,0.5,'count')

Add a slope line

see polyfit

plt.figure(figsize=(10,5))
plt.semilogy(years, process, 'o', basey=2)
plt.xlabel('year')
plt.ylabel('count')

Text(0,0.5,'count')

plt.figure(figsize=(10,5))
plt.semilogy(years, count, 'o', basey=2)
plt.xlabel('year')
plt.ylabel('count')

p = np.polyfit(years, np.log2(count), 1)
plt.plot(years, 2**(np.polyval(p, years)), 'r-')
#print(1 / p[0])

[<matplotlib.lines.Line2D at 0x110d79470>]

Make it look nice ... with Seaborn

import seaborn as sns
plt.figure(figsize=(10,5))
plt.semilogy(years, count, 'o', basey=2, ms=15)
plt.xlabel('year')
plt.ylabel('count')

p = np.polyfit(years, np.log2(count), 1)
plt.plot(years, 2**(np.polyval(p, years)), color='r', ls=':')
print(1 / p[0])

2.02342073473

increase the font, make a light background

import seaborn as sns
sns.set(font_scale=3)
sns.set_style("whitegrid") 
plt.figure(figsize=(10,5))
plt.semilogy(years, count, 'o', basey=2, ms=15)
plt.xlabel('year')
plt.ylabel('count')

p = np.polyfit(years, np.log2(count), 1)
plt.plot(years, 2**(np.polyval(p, years)), color='r', ls=':')
print(1 / p[0])

2.02342073473

Reset the tick marks

import seaborn as sns
sns.set(font_scale=3)
sns.set_style("whitegrid") 
plt.figure(figsize=(10,5))
plt.semilogy(years, count, 'o', basey=2, ms=15)
plt.xlabel('year')
plt.ylabel('count')

torig = plt.gca().get_yticks()
print(torig)
tnew = [1, 1000, np.int(1e6), np.int(1e9)]
tnewlabel = ['1', '1K', '1M', '1B']
plt.yticks(tnew, tnewlabel)


p = np.polyfit(years, np.log2(count), 1)
plt.plot(years, 2**(np.polyval(p, years)), color='r', ls=':')
print(1 / p[0])

[  8.00000000e+00   1.02400000e+03   1.31072000e+05   1.67772160e+07
   2.14748365e+09   2.74877907e+11   3.51843721e+13]
2.02342073473