CS109 Lecture 2

CS109 Lecture 2

Concepts

1. Infographics
2. Distribution
   
   • CDF (cumulative distribution function)
     python
import scipy.stats
scipy.stats.norm.cdf(2)

   • Histograms
     Histogram is easier to interpret than CDF
3. Normal Approximation
   Most data is not normal,So It’s Important for us to picturing data , looking into data and finding an appropriate way to tell a story
4. QQ-plots
   Observed versus normal approximation quantiles

Topics

Data Wrangling

In most case , we have to clean our data

Install Anconda and Use IPython

Anconda is a great environments for us to use python

Get Start With Data

import pandas as pd

1. Read data from Web
   python
url = ‘urls’
Data = pd.read_table(url)

2. Plot data
   python
data_to_plot = Data.column1
data_to_plot.plot()
data_to_plot.plot(kind='bar')
data_to_plot_multiple = Data
data_to_plot_multiple.plot(kind = 'bar')

3. Fix the Legend
   python
ax = data_to_plot.plot(kind = 'bar' , legend = False)
ax.legend(loc = 'center left' , bbox_to_anchor = (1,0.5))

4. Stacked Bar Plot
   python
data_to_plot_multiple.plot(kind = 'bar' , legend = False , stacked = True)
data_to_plot_multiple.plot(kind = 'barh' , legend = False , stacked = True) ＃ horizontal Barplot

5. Add Labels For Horizontal Barplot
   python
data_to_plot_multiple.set_index(index_names,inplace = True)
data_to_plot_multiple.plot(kind = 'barh' , legend = False , stacked = True)


Anscombe’s Quartet

Same mean,variance,correlation,and linear regression line , but the data is very different