2018-SASxSTCWorkshop

Agenda

Time	Agenda
0930	Registration start
1020	Opening
1030	Analytic Talk by Fusionex
1200	Q & A
1230	Lunch
1330	Workshop start
1620	Ending Speech
1630	End

Workshop Agenda

Time	Agenda
1330	Ice Breaking
1333	Introduce TAs
1334	Install requirement
1350	Basic Web Scrapping
1415	Scrape CIA world factbook
1445	Panda + Matplotlib
1620	End

Slides

Google Slides

Workshop

1. Install all the requirements before start

   $ pip install -r ./requirements.txt

   OR

   $ python -m pip install -r ./requirements.txt

2. Error (Windows)

   Lack of TK

   Traceback (most recent call last):
     File "<stdin>", line 1, in <module>
     File "/usr/local/lib/python3.6/site-packages/matplotlib/pyplot.py", line 115, in <module>
       _backend_mod, new_figure_manager, draw_if_interactive, _show = pylab_setup()
     File "/usr/local/lib/python3.6/site-packages/matplotlib/backends/__init__.py", line 32, in pylab_setup
       globals(),locals(),[backend_name],0)
     File "/usr/local/lib/python3.6/site-packages/matplotlib/backends/backend_tkagg.py", line 6, in <module>
       from six.moves import tkinter as Tk
     File "/usr/local/lib/python3.6/site-packages/six.py", line 92, in __get__
       result = self._resolve()
     File "/usr/local/lib/python3.6/site-packages/six.py", line 115, in _resolve
       return _import_module(self.mod)
     File "/usr/local/lib/python3.6/site-packages/six.py", line 82, in _import_module
       __import__(name)
     File "/usr/local/lib/python3.6/tkinter/__init__.py", line 36, in <module>
       import _tkinter # If this fails your Python may not be configured for Tk
   ModuleNotFoundError: No module named '_tkinter'

   Solution: Reinstall python and check TK modules

   Microsoft Visual C++ 14.0

   running build_ext
   building 'twisted.test.raiser' extension
   error: Microsoft Visual C++ 14.0 is required. Get it with "Microsoft Visual C++ Build Tools": http://landinghub.visualstudio.com/visual-cpp-build-tools

   Solution:

   1. Manually download twisted from here

   2. Run

   $ pip install Twisted‑18.9.0‑cp37‑cp37m‑win32.whl

   Lack of pywin32

   ModuleNotFoundError: No module named 'pywin32'

   Solution:

   $ pip install pywin32

   OR

   $ pip install pypiwin32

Scrapping Workshop

1. First, create a scrapy project

   $ scrapy startproject sasxstc

2. We will start by scrapping this website

3. Create a new file in projectdir/sasxstc/sasxstc/spiders/SampleSpider.py

4. Insert below boilerplate code into SampleSpider.py

   import scrapy
   
   
   class SampleSpider(scrapy.Spider):
       name = "sample"
       start_urls = [
           'https://sunwaytechclub.github.io/2018-SASxSTCWorkshop/1.html'
       ]
   
       '''
       Called after every request
       This is where your scrapping code should be
       '''
       def parse(self, response):
           result = response.body
           return {"result": result}

5. Run the spider

   $ scrapy crawl sample

6. If you get the following error:

     File "/home/gaara/.virtualenvs/sasxstc/lib/python3.7/site-packages/twisted/conch/manhole.py", line 154
       def write(self, data, async=False):
                                 ^
   SyntaxError: invalid syntax

   run

   $ pip install git+https://github.com/twisted/twisted.git@trunk

7. Xpath basic

   # Extract every html tag
   result = response.xpath('//html').extract()

   # Extract the body tag which follow by html
   result = response.xpath('//html/body').extract()

   # Extract every body tag
   result = response.xpath('//body').extract()

   # Extract the p tag that follow by div, body and html
   result = response.xpath('//html/body/div/p').extract()

   # Extract the text within every p tag
   result = response.xpath('//p/text()').extract()

8. Xpath with id and class

   # Extract every element within div with id of 'd01'
   result = response.xpath('//div[@id="d01"]').extract()

   # Extract all the text with class blue
   result = response.xpath('//p[@class="blue"]/text()').extract()

   # Extract the word SAS
   result = response.xpath('//div[@id="d02"]/p[@class="red"]/text()').extract()

Scrapping CIA factbook

CIA is Central Intelligence Agency. Various data can be found in CIA factbook, such as country GDP, population growth rate, etc. The list of data can be found in here.

1. So, first, visit this url and observe the website.

2. Let's create an empty spider FactbookSpider.py

   from pprint import pprint
   import scrapy
   
   
   class FactbookSpider(scrapy.Spider):
       name = "factbook"
       start_urls = [
           'https://www.cia.gov/library/publications/the-world-factbook/rankorder/rankorderguide.html'
       ]
   
       '''
       Called after every request
       This is where your scrapping code should be
       '''
   
       def parse(self, response):
           pass

3. Time to get the link

   links = response.xpath('//body//div[@id="profileguide"]/div[@class="answer"]//a')
   for index, link in enumerate(links):
       text = link.xpath('text()').extract_first()
       link = link.xpath('@href').extract_first()
       print(text)
       print(link)

4. Now, join the url

   link = response.urljoin(link.xpath('@href').extract_first())

5. Put the links into results object

   links = response.xpath('//body//div[@id="profileguide"]/div[@class="answer"]//a')
   results = {}
   for index, link in enumerate(links):
       text = link.xpath('text()').extract_first()
       link = response.urljoin(link.xpath('@href').extract_first())
       results[text] = link
   
   pprint(results)

6. Crawl into one of the link

   def parse(self, response):
       links = response.xpath('//body//div[@id="profileguide"]/div[@class="answer"]//a')
       results = {}
       for index, link in enumerate(links):
           text = link.xpath('text()').extract_first()
           link = response.urljoin(link.xpath('@href').extract_first())
           results[text] = link
   
       yield scrapy.Request(
           results["Population growth rate:"],
           callback=self.parse_population,
           meta={"links": results}
       )
   
   def parse_population(self, response):
       meta = response.meta
       pprint(meta)

7. Scrape the row and store into results

   rows = response.xpath('//div[@class="wfb-text-box"]//table[@id="rankOrder"]/tbody/tr')
   results = {}
   for index, row in enumerate(rows):
       if not row.xpath('@class').extract_first() == "rankHeading":
           id = row.xpath('@id').extract_first()
           name = row.xpath('td[@class="region"]//text()').extract_first()
           population_growth = row.xpath('td[3]/text()').extract_first()
           print(id + " " + name + " " + population_growth)
           results[id] = {
               "name": name,
               "population_growth_rate": population_growth
           }

8. Do the same to extract gdp growth rate

   def parse_population(self, response):
       meta = response.meta
       rows = response.xpath('//div[@class="wfb-text-box"]//table[@id="rankOrder"]/tbody/tr')
       results = {}
       for index, row in enumerate(rows):
           if not row.xpath('@class').extract_first() == "rankHeading":
               id = row.xpath('@id').extract_first()
               name = row.xpath('td[@class="region"]//text()').extract_first()
               population_growth = row.xpath('td[3]/text()').extract_first()
               results[id] = {
                   "name": name,
                   "population_growth_rate": population_growth
               }
       meta["results"] = results
       yield scrapy.Request(
           meta["links"]["Infant mortality rate:"],
           callback=self.parse_infant_mortality,
           meta=meta
       )
   
   def parse_infant_mortality(self, response):
       meta = response.meta
       results = meta["results"]
   
       rows = response.xpath('//div[@class="wfb-text-box"]//table[@id="rankOrder"]/tbody/tr')
       for index, row in enumerate(rows):
           if not row.xpath('@class').extract_first() == 'rankHeading':
               id = row.xpath('@id').extract_first()
               infant_mortality_rate = row.xpath('td[3]/text()').extract_first()
               results[id]["infant_mortality_rate"] = infant_mortality_rate
   
       return results

9. Getting error?

   Traceback (most recent call last):
     File "/home/gaara/.virtualenvs/sasxstc/lib/python3.7/site-packages/twisted/internet/defer.py", line 654, in _runCallbacks
       current.result = callback(current.result, *args, **kw)
     File "/home/gaara/Desktop/2018-SASxSTCWorkshop/sasxstc/sasxstc/spiders/FactbookSpider.py", line 60, in parse_gdp
       results[id]["infant_mortality_rate"] = infant_mortality_rate
   KeyError: 'kv'

   Surround with try and except

   try:
       results[id]["infant_mortality_rate"] = infant_mortality_rate
   except KeyError:
       pass

10. Now you have the data of factbook with you!

11. However, before we move forward, let's make use of the pipelines

    Add a field to projectdir/sasxstc/sasxstc/items.py

    class SasxstcItem(scrapy.Item):
        # define the fields for your item here like:
        results = scrapy.Field()

    Uncomment the following line in projectdir/sasxstc/sasxstc/settings.py

    # ITEM_PIPELINES = {
    #    'sasxstc.pipelines.SasxstcPipeline': 300,
    #}

    Import the item in projectdir/sasxstc/sasxstc/spiders/FactbookSpider.py

    from sasxstc.items import SasxstcItem

    Change the last line of projectdir/sasxstc/sasxstc/spiders/FactbookSpider.py

    # return results
    item = SasxstcItem()
    item["results"] = results
    return item

12. You are good to go now!

Analytics

1. Go to projectdir/sasxstc/sasxstc/pipelines.py

2. Add imports

   import pandas
   import seaborn
   from matplotlib import pyplot
   from pprint import pprint
   from scipy import stats

3. Seperate results into different list

   results = item["results"]
       country_name = []
       population_growth = []
       infant_mortality = []
   
       for country_code in list(results.keys()):
   
           country_name.append(results[country_code]["name"])
   
           population_growth.append(float(results[country_code]["population_growth_rate"]))
   
           try:
               infant_mortality.append(float(results[country_code]["infant_mortality_rate"]))
           except KeyError:
               infant_mortality.append(None)

4. Put data into Panda dataframe

   data = pandas.DataFrame(
           {
               "infant_mortality": infant_mortality,
               "population_growth": population_growth
           },
           index=country_name
       )
   
   pprint(data)

5. Run it and see how the data looks like

6. Drop the row with empty field

   data = data.dropna(how='any')
   pprint(data)

7. Plot the graph

   seaborn.jointplot(x="infant_mortality", y="population_growth", data=data, kind="reg")
   
   pyplot.show()

8. Run it!

9. Now, add R and P value?

   seaborn.jointplot(x="infant_mortality", y="population_growth", data=data,
                     kind="reg", stat_func=stats.pearsonr)

10. Add the equation of regression line

    slope, intercept, r_value, p_value, std_err = stats.linregress(data["infant_mortality"].tolist(), data["population_growth"].tolist())
    
    seaborn.jointplot(x="infant_mortality", y="population_growth", data=data,
                      kind="reg", stat_func=stats.pearsonr)
    pyplot.annotate("y={0:.1f}x+{1:.1f}".format(slope, intercept), xy=(0.05, 0.95), xycoords='axes fraction')
    pyplot.show()

11. And, you are done!