rodrigosantosbr

[Py] Tabela UF brasileiras do Wikipedia pro Excel

Mar 21st, 2019
100
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
Python 0.80 KB | None | 0 0
  1. import requests
  2. from bs4 import BeautifulSoup
  3. import pandas as pd
  4.  
  5. website_url = requests.get('https://pt.wikipedia.org/wiki/Unidades_federativas_do_Brasil').text
  6. soup = BeautifulSoup(website_url, "lxml")
  7. #print(soup.prettify())
  8. table = soup.find('table',{'class':'wikitable'})
  9. table_rows = table.find_all('tr')
  10.  
  11. res = []
  12. for tr in table_rows:
  13.     td = tr.find_all('td')
  14.     row = [tr.text.strip() for tr in td if tr.text.strip()]
  15.     if row:
  16.         res.append(row)
  17.  
  18. df = pd.DataFrame(res, columns=["UF", "Abreviação", "Capital", "Área (km2)", "População (2014)", "Densidade (2005)", "PIB (2015)", "% total 2015", "PIB per capita", "IDH (2010)", "Alfabetização (2016)", "Mortalidade infantil (2016)", "Expectativa de vida (2016)"])
  19. print(df)
  20. df.to_excel("ufbrl_wikipedia.xlsx", encoding='utf-8')
Advertisement
Add Comment
Please, Sign In to add comment