Start Jupyter Lab
- client server architecture
- browser connects to server
- kernels execute commands
- different kernels available
- terminals, text editor, notebook
You can arrange panels as you like them.
In [4]:
pwd
Out[4]:
Notebook blends text (markdown) and code (python, in our case)
- click in a box to activate it
- esc gives you access to the notebook commands (menu, etc)
- m makes a book markdown, y makes it code -- or use the menu at the top
Latex¶
For those who know TeX, the notebook can render math:
$$\sum_{i=1}^{\infty}\frac{1}{n^2} = \frac{\pi^2}{6}$$Python¶
Python is an interpreted language and the blocks of the notebook can be used as a calculator (a fancy one!)
Variables¶
In [9]:
name = "Jeremy"
age = 60
probability = .35
In [10]:
print(name, age, probability)
In [11]:
print(name, 'is', age,'years old with probability',probability)
Variables must be created before being used, by having a value assigned to them.
In [12]:
temperature
Out[12]:
In [7]:
temperature=50
In [13]:
temperature*2
Out[13]:
In [14]:
print("the temperature is:", temperature)
In [15]:
full_name = "Jeremy Teitelbaum"
In [16]:
full_name[3]
Out[16]:
In python, indexing starts from zero!!!
In [17]:
full_name[0]
Out[17]:
In [18]:
len(full_name)
Out[18]:
Types¶
- strings
- integers
- floats
- lists
In [25]:
a=123
b='123'
In [26]:
a[1]
In [27]:
b[1]
Out[27]:
In [512]:
L = ['jeremy', 'kendra', 'parkisheet']
In [513]:
L[0]
Out[513]:
Slicing¶
In [19]:
full_name[1:3]
Out[19]:
In [20]:
full_name[5:]
Out[20]:
In [21]:
full_name[:5]
Out[21]:
In [22]:
full_name[3:5:3]
Out[22]:
In [23]:
full_name[:-1]
Out[23]:
In [24]:
full_name[-1]
Out[24]:
In [28]:
full_name[-1::-1]
Out[28]:
In [514]:
L = ['a','b','c','d']
In [515]:
L[1:3]
Out[515]:
In [516]:
L[-1]
Out[516]:
In [517]:
L[::2]
Out[517]:
In [29]:
magic_word = "thanos_must_die"
- What command will determine how many letters are in the
magic_word? - How would you print the fifth through eighth letters of the magic word (inclusive)?
In [31]:
print(magic_word[4:8])
print(magic_word[7:])
print(magic_word[:3])
print(magic_word[:])
print(magic_word[2:-3])
print(magic_word[-3:2:-1])
print(magic_word[0:100])
Types¶
In python, every variable has a type, but the language figures out the appropriate type by itself.
The most important types are:
- strings
- integers
- floating point
In [32]:
x="this is a string"
a=137.50
b=12e-1
c=133
In [33]:
print(a)
In [34]:
print(b)
In [35]:
print(c)
In [37]:
type(a)
Out[37]:
In [38]:
type(b)
Out[38]:
In [39]:
type(c)
Out[39]:
In [40]:
type(x)
Out[40]:
Operations on numbers and strings¶
They're what you'd expect, with some caveats.
In [44]:
x=3.5
y=x/5
z=x*22.4
w=x*1e-1
print("x=",x,"y=",y,"z=",z,"w=",w)
String addition is concatenation.
In [45]:
x="Jeremy"+"Teitelbaum"
In [46]:
print(x)
Mixing floats and integers is ok (you get a float) but mixing strings and numbers is a problem.
In [50]:
1+3.5
Out[50]:
In [51]:
"Jeremy"+3
You can convert floats or ints to strings and then combine them:
In [55]:
"Jeremy Teitelbaum is "+60+" years old"
In [56]:
"Jeremy Teitelbaum is "+str(60)+" years old"
Out[56]:
In [58]:
3*"a"
Out[58]:
In [59]:
type("hello")
Out[59]:
In [60]:
type(60)
Out[60]:
In [61]:
type(3.6)
Out[61]:
In [62]:
len("Jeremy")
Out[62]:
In [63]:
len(3.5)
Operations on lists¶
In [518]:
L = ['jeremy','kendra', 'pariksheet','dyanna']
In [519]:
S = ['mouse', 'cat']
In [520]:
L+S
Out[520]:
In [521]:
sorted(L)
Out[521]:
In [526]:
['a']*3
Out[526]:
In [525]:
L = L + ['x']
print(L)
strings to lists and back¶
In [549]:
list('abcdefg')
Out[549]:
In [551]:
''.join(['a','b','c'])
Out[551]:
Dictionaries¶
In [283]:
f = {}
f['jeremy']=15
f['kendra']=33
f['marshmallow']=100
f['french_fries']='hello'
In [284]:
f['jeremy']
Out[284]:
In [285]:
f['french_fries']
Out[285]:
In [286]:
f[0]
In [287]:
f.keys()
Out[287]:
In [288]:
f.values()
Out[288]:
In [289]:
a = dict(name=['jeremy','kendra','pariksheet'],gender=['m','f','m'])
In [279]:
a
Out[279]:
Comments¶
In [64]:
# this goes in a code cell, but doesn't get executed
Built-in Functions¶
f(x,y,z) returns a value¶
- len
- int, str, float
- max
- min
- round
- help - function or shift-Tab
In [66]:
round(3.12131,3)
Out[66]:
In [71]:
max('hello')
Out[71]:
In [72]:
min("hello")
Out[72]:
In [73]:
round("hello")
In [74]:
1/2
Out[74]:
In [75]:
result = 60
In [77]:
report = "the result is "+str(result)
In [78]:
print(report)
In [465]:
#help(type)
In [80]:
help(round)
In [81]:
round(123,-2)
Out[81]:
In [82]:
round(1351,-1)
Out[82]:
In [466]:
#help()
In [467]:
report = "Now is the time to flee'
In [468]:
1/0
In [469]:
y = 15-23
x = y+8
z= 3/x
In [470]:
max()
In [471]:
min()
Libraries¶
Python is a very simple language¶
- There are only 33 "keywords" in the python language
False, class, finally, is, return, None, continue, for, lambda, try, True, def, from, nonlocal, while, and, del, global, not, with, as, elif, if, or, yield, assert, else, import, pass, break, except, in, raise
All of the interesting capabilities of the language come from extensions to the basic language; these extensions are called "libraries" or "modules".
The language also includes about 100 built-in functions. The functions
int, len, print, str, type, max, min
are built-in library. The built-in functions are very primitive.
To get access to more interesting functions, you need to
import
them by importing the library that defines them. Many interesting capabilities are added by the standard library
Two important libraries¶
- numpy
- pandas
In [529]:
import numpy
import pandas
In [531]:
print('pi is ',numpy.pi)
In [532]:
print('e is ',numpy.exp(1))
Notice that we refer to the function exp, for example, by first naming the library it came from.
In [533]:
exp(1)
Out[533]:
In [94]:
math.exp(1)
Out[94]:
You can get help on a library/module using the help command.
The import command also allows you to adopt abbreviations.
In [534]:
import numpy as np
In [535]:
np.exp(1)
Out[535]:
In [536]:
from numpy import exp
In [537]:
exp(1)
Out[537]:
In [538]:
import numpy.random
In [543]:
numpy.random.choice([1,2,3,4,5])
Out[543]:
In [554]:
np.random.choice(list('ACTTGCTTGAC'))
Out[554]:
In [556]:
import math
import random
bases = "ACTTGCTTGAC"
n_bases = len(bases)
idx = np.random.randint(n_bases)
print("random base", bases[idx], "base index", idx)
- Data frames are "like" spreadsheets
- Columns
- Index
- manipulation is through "methods"
In [557]:
import numpy.random as rnd
In [560]:
rnd.randint(20)
Out[560]:
Numpy arrays¶
In [700]:
a=np.arange(0,10,.1)
In [702]:
numpy.cos(a)
Out[702]:
In [703]:
3*a
Out[703]:
In [704]:
a*a
Out[704]:
In [705]:
a+numpy.exp(a)
Out[705]:
DataFrames and Pandas¶
In [108]:
import pandas as pd
simple dataframes and dictionaries¶
In [280]:
df = pd.DataFrame.from_dict(a)
In [281]:
df
Out[281]:
In [562]:
data = pd.read_csv('../gapminder_data.csv')
In [563]:
data.columns
Out[563]:
In [564]:
data.head()
Out[564]:
In [565]:
data = pd.read_csv('../gapminder_data.csv', index_col='country')
In [566]:
data.columns
Out[566]:
In [567]:
data.info()
selecting elements from dataframes¶
columns¶
In [572]:
data['pop']
Out[572]:
In [573]:
data[['year','pop']]
Out[573]:
rows¶
from the index¶
In [574]:
data.loc['Afghanistan']
Out[574]:
boolean selection on a column¶
In [575]:
data['continent']=='Asia'
Out[575]:
In [576]:
data[data['continent']=='Asia']
Out[576]:
In [618]:
data[data['pop']>1e8]
Out[618]:
In [619]:
data[data['lifeExp']<60]
Out[619]:
numerical indexing¶
In [620]:
data.iloc[33]
Out[620]:
In [621]:
data.iloc[33,3]
Out[621]:
rows and columns¶
In [577]:
data.loc['Afghanistan','pop']
Out[577]:
In [ ]:
statistics on numerical columns¶
In [579]:
data.loc['Afghanistan','pop'].describe()
Out[579]:
Grouping: optional¶
In [625]:
s=data.groupby(['continent']).mean()
s.head()
Out[625]:
In [632]:
averages_by_continent=data.groupby(['continent','year']).mean()
averages_by_continent.loc['Asia'].round()
Out[632]:
Pivot table¶
In [652]:
lifeExp_over_time = pd.pivot_table(data,index='country',columns='year',values='lifeExp')
In [653]:
lifeExp_over_time.loc['Afghanistan',:]
Out[653]:
In [656]:
stats_over_time = pd.pivot_table(data,index=['continent','country'],columns='year',values=['lifeExp','gdpPercap','pop'])
stats_over_time_by_continent=stats_over_time.groupby('continent').mean()
In [660]:
stats_over_time_by_continent['pop'].round(-3)
Out[660]:
Plotting¶
plotting libraries:¶
In [662]:
import matplotlib.pyplot as plt
plt.style.use('ggplot')
In [672]:
plt.plot([1,2,3],[1,4,9],c='blue',linewidth=3,linestyle='solid')
plt.xlabel('x axis')
plt.ylabel('y axis')
plt.suptitle('A demo plot')
Out[672]:
In [675]:
plt.plot([1,2,3],[1,4,9],c='orange',linewidth=1,linestyle='dashed')
plt.xlabel('x axis')
plt.ylabel('y axis')
plt.suptitle('A demo plot')
Out[675]:
In [699]:
plt.scatter([1,2,3],[1,4,9],c=[0,1,2],s=100)
plt.plot([1,2,3],[1,4,9])
plt.xticks([1,2,3])
plt.yticks([1,4,9])
plt.xlim(0,5)
plt.ylim(0,10)
plt.grid(False)
In [ ]:
Plotting from pandas¶
- the index is the x axis and the values in the column are plotted against that
- or you can specify x and y
In [706]:
data.head()
Out[706]:
In [707]:
lifeExp_over_time.head()
Out[707]:
In [737]:
lifeExp_over_time.T['Afghanistan'].plot()
Out[737]:
In [713]:
transpose = lifeExp_over_time.T
transpose[['Afghanistan','Germany']].plot(kind='bar')
Out[713]:
In [753]:
_=stats_over_time.groupby('continent').mean()['gdpPercap'].T.plot()
In [751]:
stats_over_time.groupby('continent').mean()['lifeExp'].T.plot()
_=plt.suptitle('Mean Life Expectancy over Time')
In [754]:
_=stats_over_time['gdpPercap'].loc[['Africa']].mean().T.plot(kind='bar',legend=None)
In [774]:
data.plot(kind='scatter',y='gdpPercap',x='lifeExp',c='year',logy=True,figsize=(10,10))
plt.savefig('scatter.png')
In [763]:
pwd
Out[763]:
In [783]:
_=data[data['year']==2002].groupby('continent').sum()['pop'].plot(kind='pie',figsize=(10,10),label='Population')
In [805]:
_=data[(data['year']==2002) & (data['continent']=='Asia')].groupby('country').sum()['pop'].sort_values().plot(kind='bar',figsize=(10,10),)
In [ ]: