#!/usr/bin/env python
# coding: utf-8

# ## Corona Virus analysis using 'Novel Corona Virus 2019-20 Dataset'

# As Github does not support folium map and some rendering techniques, so if you want to see fully rendered notebook, click on below link
# 
# https://nbviewer.jupyter.org/github/Mr-Piyush-Kumar/Data_Science_Projects/blob/master/Corona_virus_analysis/CoronaVirusAnalysis.ipynb

# ### 1. Introduction

# ### 1.1 Background
# 
# From World Health Organization - On 31 December 2019, WHO was alerted to several cases of pneumonia in Wuhan City, Hubei   Province of China. The virus did not match any other known virus. This raised concern because when a virus is new, we do not know how it affects people.
# 
# So daily level information on the affected people can give some interesting insights when it is made available to the broader data science community.

# ### 2. Data acquisition and cleaning

# All dataset will be download from below Kaggle webpages. 
#                          
# Corona Virus dataset <a href = 'https://www.kaggle.com/sudalairajkumar/novel-corona-virus-2019-dataset'>from here</a>  
# World Coordinates dataset <a href = 'https://www.kaggle.com/parulpandey/world-coordinates'>from here</a>  
# World geoJson file <a href = 'https://www.kaggle.com/sandeep04201988/worldcountries'>from here</a>  
# China geoJson file <a href = 'https://www.kaggle.com/gpreda/china-regions-map'>from here</a>

# In[1]:


# Downlodaing Corona Virus Dataset

get_ipython().system('pip install wget')

import wget
url = 'https://storage.googleapis.com/kagglesdsdata/datasets/494724/966440/2019_nCoV_data.csv?GoogleAccessId=web-data@kaggle-161607.iam.gserviceaccount.com&Expires=1582868558&Signature=JXjN269XVMhWbNH3BAyQozqwNFBePHHrrxpRMk%2BpD3ANojzrkvE5Qer9z1fk1uprgM0njfohll0yPHB8mZs4eN0Ucke7I4JtTTRyISKFWSO%2FCcDZanS9pWX%2FdnEBUV6QindIhmE%2FZeeeRAeKx%2Fd8h3RGlHpCeBm6%2BN7L6XnZj8N%2FsxlmIz%2FXwbxc%2BW2VRi4pG1P0AmJI5PjIc5OVcQISdy0RCGDaJuhAGm4TTarYuZt9j36rFlheXNXitn2EeutI4OI%2B%2BBqTV7fIR5mdBXCa1%2FKiutt3Lu%2BffT7N62dlMXLR4Z0aga4OvzzzwdVuf7TbLWBpqcx6LTQGGD03Fj5kEw%3D%3D&response-content-disposition=attachment%3B+filename%3D2019_nCoV_data.csv'
file = wget.download(url)
print('file name = ',file)


# In[2]:


# Downloading word-coordinates data for visulization on world map

url1 = 'https://storage.googleapis.com/kagglesdsdata/datasets/500356/927175/world_coordinates.csv?GoogleAccessId=web-data@kaggle-161607.iam.gserviceaccount.com&Expires=1582868625&Signature=soCAku6MmJLB%2FL48Oxs%2FSxaPP11EASBJG%2BPMCRfhPcavNJBDhviiE12SKcfAje07vuDsWsuRiUxRu61KgBObNNe2brA9JHiEL4I4D1dgYk%2F6SKLJz6RFWCwYnr5b%2FAiHfQW2dLxNtRF35MEfR4ncnOW0FHLzgSTd8wq12QrV6y9%2B8XAOyjwcFdkCnLk0eTws0ERhUMC0CSeqFXqkS7leuBxSAUW2jUblrLvir3938JSZ2aqqbywj4eIT5jy0XvmpxepDsYOu1kJpkkyKq4lzyWTwNmr5ljZGWSoJrOVMAIU17z%2FnI3Is2s2YyysoUI6I0XoXRPqyN%2B8pZy0%2F0GPw1Q%3D%3D&response-content-disposition=attachment%3B+filename%3Dworld_coordinates.csv'
file1 = wget.download(url1)
print('file name = ',file1)


# In[3]:


# downloading world geoJson file to draw choropleth map of world

url2 = 'https://storage.googleapis.com/kagglesdsdata/datasets/7923/11172/world-countries.json?GoogleAccessId=web-data@kaggle-161607.iam.gserviceaccount.com&Expires=1582868666&Signature=ca8CoGxJ1mYdOR%2BKnaYdfO5mW15iNPvFWNOElOCDMUMRAU4wfO9G2W5980D6zYDFO94SrgiFm07enAWeuF3SmUF5XuU70bn%2FDID%2F%2F2%2FU4JHE00szhJWsIyLHm66R%2FInBNR4Wqg1OYgA0qlqw6Ewy9%2F0jAl4P%2Bv62voX7oO1rI%2BEGPwS0DBaHsUHCGwL4GNdxjJJ%2Ba05c6jpksTvP4V%2FWk%2FvLeyGOMa%2BE5FzExHInBlbiCCPX5bsUo%2BjG87kpIokoSIt2eoP7fiVzPnhB6Z83Td3zwr5ld0gJA%2FyxXagrDbB70JwmdgVO%2B3k2%2BySZqgev%2F6rLk7Lm%2FJFy%2B6ZMgPIVcw%3D%3D&response-content-disposition=attachment%3B+filename%3Dworld-countries.json'
file2 = wget.download(url2)
print('file name = ',file2)


# In[4]:


# downloading china geoJson file to draw choropleth map of China

url3 = 'https://storage.googleapis.com/kagglesdsdata/datasets/496669/922532/china.json?GoogleAccessId=web-data@kaggle-161607.iam.gserviceaccount.com&Expires=1582868706&Signature=L4Xo4fynYwWPvUio3j1NWc4sQBFeAje1N3kZ4%2BzrLJdFZsQiyMJB3lisf%2FVS%2BD7Kq0Pe0aCgxhS1tpmaTbQFF87255Fy%2B7In%2BJOJZnmytgc8NGc6m9ZaOlzuLkgoAT3tgfLYWmRcWLvvBl1cdAArfhl%2BibxSl7%2BbjNvGo3Q0XolnhfdjKZCZ5egh%2FmSV8HUfrf0OBm0%2BXbUEcmV4wkJm8RM8UFSvE1t5HFrhvzO29WmZm9rcwiHKuLPxjl3KGtXpp4ysTJPpScM50UyAgWcrZH1LYeARmGhbnGRarblLGfeHoj2sbO9Nh%2FQAd5XmiNEMtdxTnR99p65Dd07v76979Q%3D%3D&response-content-disposition=attachment%3B+filename%3Dchina.json'
file3 = wget.download(url3)
print('file name = ',file3)


# In[5]:


# loading corona virus dataset as Pandas Dataframe
import pandas as pd

corona_df = pd.read_csv('2019_nCoV_data.csv')
corona_df.head()


# **Corona Dataset columns decription:-**
# 
# - Sno:- Serial Number.  
# - Date:- In which date and time, a particular record wrote in the dataset.  
# - Province/State:- Province/State name of belonging country.  
# - Country:- Country Name.  
# - Last Update:- At what time, last case was observed in a particular day.  
# - Confirmed:- No. of confirmed infected people till now in a particular  location (Representing Cumulative Sum).  
# - Deaths:- No. of deaths till now due to Corona virus in a particular  location (Representing Cumulative Sum).  
# - Recovered:- No. of recovered people till now who was infected in a particular  location (Representing Cumulative Sum).  

# In[6]:


# loading both geoJson files

world_geoJson = 'world-countries.json' # Path to file
china_geoJson = 'china.json' # Path to file


# **About geoJson file**
# 
# GeoJSON is an open standard geospatial data interchange format that represents simple geographic features and their nonspatial attributes. Based on JavaScript Object Notation (JSON), GeoJSON is a format for encoding a variety of geographic data structures. It uses a geographic coordinate reference system, World Geodetic System 1984, and units of decimal degrees.
# 
# <a href = 'https://doc.arcgis.com/en/arcgis-online/reference/geojson.htm'>Read more about geoJson file from here.</a>
# 
# I will use these two files to show choropleth map of world as well as of china to representing corona virus outbreak location wise.
# 
# <a href='https://en.wikipedia.org/wiki/Choropleth_map'>Read about choropleth map from here.</a>

# ### Data Cleaning

# dropping Sno. from Dataset as it's not required for analysis.

# In[7]:


corona_df.drop('Sno',axis=1,inplace=True)
corona_df.head()


# In[8]:


corona_df.Country.unique()


# China and Mainland China are representing as two different countries in dataset. Lets replace Mainland china with China.

# In[9]:


corona_df.replace('Mainland China','China',inplace=True)
corona_df.Country.unique()


# In Country column, there is one value 'Others'. As I have no information about how many countries are grouped as Others so I am going to drop those rows from dataset having country value Others.

# In[10]:


corona_df = corona_df[corona_df['Country']!='Others'].reset_index(drop=True)
corona_df


# ### 3. Data Analysis

# **Handling NaN values**

# In[11]:


# shape of our dataset
corona_df.shape


# total no of 1719 records are there in corona dataset.

# In[12]:


# looking for null values present in corona datset

corona_df.isna().sum()


# In Province/State columns total 462 records have null value.

# In[13]:


import numpy as np

corona_province_nan = corona_df[corona_df['Province/State'].isna()]
corona_province_nan


# In[14]:


corona_province_nan.Confirmed.unique()


# So the conclusion is that if a country has low no. of cases then dataset creators didn't divide country data into Province/State.

# Replacing Nan Values with respective country name.

# In[15]:


pd.options.mode.chained_assignment = None # avoiding setting with copy warning
corona_df['Province/State'][corona_df['Province/State'].isna()] = corona_df['Country']


# In[16]:


print('Total',len(corona_df.Country.unique()),'countries have been affected by corona virus till now.')


# In[17]:


# total cases till now country wise

countrywise_df = corona_df.groupby(['Country','Province/State']).max()
countrywise_df.reset_index(inplace = True)
countrywise_df = countrywise_df.groupby('Country').sum()
countrywise_df.reset_index(inplace = True)
countrywise_df


# China is most affected by Corona virus.

# In[18]:


import seaborn as sns
import matplotlib.pyplot as plt


# ### **Visualizing spreading of corona virus from beginning to till now. day by day** 
# 
# as we have per day data of each country so there is no problem in it.

# In[19]:


Day_wise_world_data = corona_df.groupby('Date').sum().reset_index()
Day_wise_world_data


# In[20]:


Day_wise_china_data = corona_df[corona_df['Country']=='China'].groupby('Date').sum().reset_index()
Day_wise_china_data


# In[21]:


# World
plt.figure(figsize=(20,10))
plt.subplot(1,2,1)
plt.plot(Day_wise_world_data.Date,Day_wise_world_data.Confirmed,label='Confirmed',color='r',marker='o')
plt.plot(Day_wise_world_data.Date,Day_wise_world_data.Recovered,label='Recovered',color='g',marker='o')
plt.plot(Day_wise_world_data.Date,Day_wise_world_data.Deaths,label='Deaths',color='b',marker='o')
plt.title('World',size=14)
plt.legend()
plt.xticks(rotation=90)
plt.xlabel('Date',size=14)
plt.ylabel('No. of cases',size=14)

# China
plt.subplot(1,2,2)
plt.plot(Day_wise_china_data.Date,Day_wise_china_data.Confirmed,label='Confirmed',color='r',marker='o')
plt.plot(Day_wise_china_data.Date,Day_wise_china_data.Recovered,label='Recovered',color='g',marker='o')
plt.plot(Day_wise_china_data.Date,Day_wise_china_data.Deaths,label='Deaths',color='b',marker='o')
plt.title('China',size=14)
plt.legend()
plt.xticks(rotation=90)
plt.xlabel('Date',size=14)
plt.ylabel('No. of cases',size=14)
plt.suptitle('No. of cases of Corona virus day by day',size=15)
plt.show()


# The pattern of spreading of corona virus is same for both whole world and China, it's because mostly Chinese people are affected by corona virus and whole world data is totally depend on what is happening in China.
# 
# So we can find hidden pattern in Corona virus dataset using only Chinese Data.

# No. of confirmed cases are increasing exponentially but recovery is not fast as it should be.

# ### **Analyzing China Data State wise**

# In[22]:


China_data = corona_df[corona_df['Country']=='China']
China_data.head()


# In[23]:


China_data_state_wise = China_data.groupby('Province/State').max().reset_index()
China_data_state_wise.head()


# In[24]:


plt.figure(figsize=(15,12))
ax = sns.barplot(x='Confirmed',y='Province/State',data=China_data_state_wise,errwidth=0)
plt.title('No. of Cases in China till now (State wise)',size=15)
plt.xlabel('No. of Confirmed cases',size=14)
plt.ylabel('Province/State',size=14)

y=0
for p in ax.patches :
   ax.annotate(str(int(p.get_width())),(p.get_width()+100,y+0.1),color='b')
   y+=1


# Hubei province is most affected by Corona virus in China, as corona virus started from Wuhan city which is in Hubei province.

# In[25]:


Country_wise_except_china = countrywise_df[countrywise_df['Country']!='China']
Country_wise_except_china.head()


# In[26]:


plt.figure(figsize=(15,12))
sns.set_color_codes('pastel')
sns.barplot(x='Confirmed',y='Country',data=Country_wise_except_china,errwidth=0,color='r',label='Confirmed')
sns.barplot(x='Recovered',y='Country',data=Country_wise_except_china,errwidth=0,color='g',label='Recovered')
sns.barplot(x='Deaths',y='Country',data=Country_wise_except_china,errwidth=0,color='b',label='Deaths')
plt.title('No. of ceases till now country wise except China',size=14)
plt.xlabel('No. of cases',size=13)
plt.ylabel('Country',size=13)
plt.legend()
plt.show()


# ### **Geographical Visualization**

# For Geographical Visualization I am using Folium library

# In[27]:


get_ipython().system('pip install folium==0.10')


# In[28]:


import folium as flm


# In[29]:


# loading world coordinates file

world_coordinates = pd.read_csv('world_coordinates.csv')
world_coordinates.head()                                


# dropping code from world_coordinates Dataframe

# In[30]:


world_coordinates.drop('Code',axis=1,inplace=True)
world_coordinates.head()


# Merging world_coordinates with country wise df

# In[31]:


corona_df_with_geo_location = countrywise_df.merge(world_coordinates,how='left',on='Country')
corona_df_with_geo_location


# In[32]:


# dropping rows with NaN values
corona_df_with_geo_location.dropna(axis=0,inplace=True)
corona_df_with_geo_location


# In[33]:


world_map = flm.Map(location=[10,-10],zoom_start=2)
for country,confirmed,lat,lng in zip(corona_df_with_geo_location.Country,
                                     corona_df_with_geo_location.Confirmed,
                                     corona_df_with_geo_location.latitude,
                                     corona_df_with_geo_location.longitude):
    
    flm.CircleMarker(
                      location=[lat,lng],
                      radius=5,
                      tooltip=str(confirmed),
                      fill=True,
                      fill_color='red',
                      fill_opacity=0.8).add_to(world_map)
world_map


# In the above world map circle marker presents that these countries have been affected by corona virus.

# **Choropleth map**

# In[34]:


myscale = [0,15000,30000,45000,60000,73090] # manually defining threshold scale

choropleth_map = flm.Map(location=[10,-10],zoom_start=2)


flm.Choropleth(
                           geo_data = world_geoJson,
                           data = corona_df_with_geo_location,
                           columns = ['Country','Confirmed'],
                           key_on='feature.properties.name',
                           fill_color= 'YlOrRd',
                           fill_opacity=0.7, 
                           line_opacity=0.2,
                           legend_name='No. of confirmed cases',
                           threshold_scale = myscale,
                           nan_fill_color='black',
                           nan_fill_opacity=0.4,
                           highlight=True
                           ).add_to(choropleth_map)
choropleth_map


# - Black colored areas represents that there is no data for these countries in dataset.
# - Light yellow colored areas have less no of cases.
# - China is red colored because there are very large no. of cases.
# 
# for more you can refer to threshold scale on map.

# **Now visualizing the choropleth map for china only(Province wise)**

# In[35]:


# loading china geoJson file

import json

with open(china_geoJson) as file:
    china = json.load(file)


# Checking whether province/state name in china geoJson file are in chinese language or in english language.

# In[36]:


name_lst = []
for i in range(len(china['features'])):
    name_lst.append(china['features'][i]['properties']['name'])
name_lst


# Province/State name are in chinese language, need to translate in english because data present in dataset are in english language.

# In[37]:


# installing google translator library 

get_ipython().system('pip install googletrans')


# In[38]:


from googletrans import Translator # importing translator 
translator = Translator() # creating an object of translator

name_lst_en = []
for text in name_lst:
    name_lst_en.append(translator.translate(text).text)
name_lst_en


# **Comparing above list of Province/State with Province/State in corona dataset.**

# In[39]:


China_data_state_wise['Province/State']


# **Need to modify name_lst_en according to dataset.**

# In[40]:


# Keeping only first word in every item of name_lst_en

for i in range(len(name_lst_en)):
    name_lst_en[i] = name_lst_en[i].split(" ")[0]
name_lst_en


# **now translated names are matching with the names present in dataset.**

# In[41]:


# writing the tranlated name of province/state in china geoJson file

for i in range(len(name_lst_en)):
    china['features'][i]['properties']['name']=name_lst_en[i]


# In[42]:


China_lat = 35.8617
China_lng = 104.1954

China_map = flm.Map(location=[China_lat,China_lng],zoom_start=4)

flm.Choropleth(
               geo_data=china,
               data=China_data_state_wise,
               columns=['Province/State','Confirmed'],
               key_on='feature.properties.name',
               fill_color='YlOrRd',
               fill_opacity=0.7,
               nan_fill_color='black',
               nan_fill_opacity=0.7,
               highlight=True,
               legend_name='No. of confirmed cases',
                ).add_to(China_map)
China_map


# - Black colored areas represents that there is no data for these states in dataset.
# - Light yellow colored areas have less no of cases.
# 
# for more you can refer to threshold scale on map.

# Only one province is of red color, indicates mostly infected people are here. It's Hubie.

# ## Thank you for reading