00. intro to pandas

• Package pandas is for data analysis.
• Bacause of this package, you can create a dataframe object in python.
• A dataframe can be manuipulated and virtualized(plotting).

01. my data

steps
• Use my created data for this demo.
• create a file in any editor, pandas_01_mydata.py as below.
• in browser, open rept.it/languages/python3
• click ther icon for my repl, select a new.
• create a subsirectory, mydir
• upload the create py file into mydir.
• click package icon, add package pandas-datareader.
• update main.py
  print("enter main.py")
  from mydir import pandas_01_mydata
• click run
• If the image does not show, edit a bit, like print("12/04/2018 11:45") on the new py file to trigger, kind of patch.

        #dictionary to df
        import pandas as pd

        print("12/04/2018, 11:33 am")

        access_data = {'Day':[1,2,3,4,5],
                     'Visitors':[99,22,33,44,11],
                     'Bounce Rate':[15,20,30,40,50]}

        df = pd.DataFrame(access_data)


        # modify df to make Day as the index
        df.set_index('Day', inplace=True)
        print(df)

        import matplotlib.pyplot as plt
        from matplotlib import style
        style.use('fivethirtyeight')

        df.plot()
        #plt.show()
        plt.savefig('pandas_p01.png')
            

code review
• The site is repl.it/languages/python3.
• package python3, and its related packages - numpy, matplotlib are not required to prepare.
• line 6, create a dictionary object
  
  • key is a string.
  • value is a list.
• line 10, create a dataframe from the dictionary object.
• line 14, set 'Day' as the index.
• line 15, print(df), the data is presented in a tabular form.It demonstrates to use a dataframe.
• line 19, style.use('fivethirtyeight') gives a better look.
• line 21, plot
• line 22, comment out because this is on the web.
• line 23, save it as an image file instead.It demonstrate for data virtualization.

02. stock data from yahoo

steps
• Use the data from yahoo for this demo.
• create a file in any editor, pandas_02_yahoo.py as below.
• in browser, open rept.it/languages/python3
• click ther icon for my repl, select the one for the previous lab.
• upload the create py file into mydir.
• update main.py
  print("enter main.py")
  from mydir import pandas_02_yahoo
• click run
• If the image does not show, edit a bit, like print("test at5:32") on the new py file to trigger, kind of patch.
• The py code is as below:

        #pandas_02_yahoo.py
        print("test at 5:30 pm")

        import pandas as pd
        import datetime
        import pandas_datareader.data as web

        start = datetime.datetime(2015, 1, 1)
        end = datetime.datetime.now()
        
        df = web.DataReader("AAPL", "yahoo", start, end)
        
        print(df.head())

        import matplotlib.pyplot as plt
        from matplotlib import style
        style.use('fivethirtyeight')

        df['High'].plot()
        plt.legend()
        #plt.show()
        plt.savefig('pandas_02_yahoo.png')
        

• The code highlighted in red is the main code.
• pandas_datareader.data is a module.
• in its function DataReader
  • argument 1: "AAPL" is the stock symbol.
  • argument 2: data from Yahoo.
  • argument 3,4: the date range
  • return in a dataframe.
  • You can see the output and the plotting.
• This way is similar to accessing web services.
  - All are in one FUNCTION CALL.
  - It takes care of its connection and prototocals.

03. create a dataframe from a csv file

steps
• Using code visual studio in mac
• create a working folder.
• prepare a cvs file under the folder
• prepare a python file under the same folder, code as below
• run it.

        import pandas as pd
        df = pd.read_csv('mydata.csv') 
        print(df)           
            

04. installing python3.7 + pandas + other package, using code visual studio

problem background
• The context in this topic evolves fast. This topic is based on my lab on December 8, 2018.
• I worked on mac. Previously, I had python3.7 installed, and I used code visual studio.
• Previously, when I used pip installed a python package, an error was experienced MODULE NOT FOUND.
• Then, I used a service from website repl.it to develop the code for data analysis.
• If I did not involve package pandas, I still can use code visual studio to test.

my experience
• I found that a third party download, Anaconda3 python distribution for python 3.7, pandas, and other packages
• in my mac, I did not uninstall python, code visual studio.
• in my browser, open https://www.continuum.io
• select download
• mac osx64-bit, graphical installer
• click download, it took quite a while.
• There are many steps executed automatically in sequence.
• the last step is for code visual studio. It asked a question.
• Which INTERPRETER do you want CONDA or regular.
• select CONDA.
• Finally, the installation is completed.
• Then I opened code visual studio, create a new file as below.
• test the code without any error.

        import numpy
        import pandas
        import matplotlib
        print("So far so good")
            

comments
• This type installation is for pandas related apps for data analysis.
• The interpreter from a third party is REQUIRED. It is DIFFERENT from the one from python website.
• Anaconda3 python distribution installs the interpreter, python3.7, pandas, other packaged in your device. You do not need download them during developing time.
• Package is a very unque stuff for python. R language uses package also, there are some differences.

demo - using code visual studo, after Anconda3 installed

        # using w3c python tutorial example for if...elif...else

        a = 200
        b = 33
        if b > a:
          print("b is greater than a")
        elif a == b:
          print("a and b are equal")
        else:
          print("a is greater than b")
        

steps
• create an folder, peter_pandas
• open code visual studio, and open the folder.
• add a file, w3c_14_if.py, save, and run.
• You can see the result.
• In the status bar at the bottom,message:Python 3.7.0 64 bit 'base':conda shows up
• In the teerminal, you also see the code is run by /users/peter/anaconda3/bin/python.
• The test in code visual studio is done.
• open my mac terminal window.
• navigate to /Users/peterkao/anaconda3/bin
• enter python w3c_14_if.py
• you'll see the result, not from the ide.

code analysis - condtions
• if, elis, else all have   :  at the end.
• Code indent is used.
• language construct if-elif-else is applied for 3+ conditions.