import pandas as pd

import datetime

import plotly.express as px

import re

mod = pd.read_csv(‘MODERNA.csv’,parse_dates=[‘EV Gateway Receipt Date’])

pfi = pd.read_csv(‘PFIZER.csv’,parse_dates=[‘EV Gateway Receipt Date’])

ast = pd.read_csv(‘ASTRA.csv’,parse_dates=[‘EV Gateway Receipt Date’])

jan = pd.read_csv(‘JANS.csv’,parse_dates=[‘EV Gateway Receipt Date’])

len(mod),len(pfi),len(pfi2),len(ast),len(jan)

mod.head()

pd.set_option(‘display.max_colwidth’, 1500)

mod[‘Reaction List PT (Duration – Outcome – Seriousness Criteria)’]

mod[‘Suspect/interacting Drug List (Drug Char – Indication PT – Action taken – [Duration – Dose – Route])’]

mod[‘VAXX’] = ‘MODERNA’

pfi[‘VAXX’] = ‘PFIZER’

jan[‘VAXX’] = ‘JANSSEN’

ast[‘VAXX’] = ‘ASTRAZENECA’

eucovid  = pd.concat([mod,pfi,ast,jan],ignore_index=True,axis=0)

eucovid.columns

len(mod) + len(pfi) + len(ast) + len(jan)

z = len(mod) + len(pfi) + len(ast) + len(jan)

eucovid19= eucovid.drop(eucovid.columns[[1,3,4,5,7,8,9,11,12,13]],axis=1).copy(deep=True)

eucovid19

eu= pd.DataFrame()

eu[[‘ID’,’Date’,’Age’,’Reaction’,’Vaccine’]]= eucovid19[[‘EU Local Number’,’EV Gateway Receipt Date’,’Patient Age Group’,’Reaction List PT (Duration – Outcome – Seriousness Criteria)’,’VAXX’]]

eu = eu.apply(lambda x: x.astype(str).str.upper())

eut = eu.copy(deep=True)

eut

regex1 = (r’\([^()]*\)’ )

regex2 = (r'<BR>’)

eut[‘Reaction’] = eu.Reaction.str.replace(regex1,”).str.replace(regex2,”)

eut

eut[‘Reaction’] = eut[‘Reaction’].str.split(‘,’)

eux = eut.explode(‘Reaction’)

eux

eux[‘Reaction’].value_counts()

eur = pd.DataFrame()

eur.index.set_names(‘REACTION’,inplace=True)

eur[‘Total_Count’]=eux[‘Reaction’].value_counts()

eur

example = eur.query(‘Total_Count > 10000’)

example

fig= px.bar(example, y=’Total_Count’,width=1500, height=900, color=example.index, title=f”Serious Reactions Reported {z}”)

fig.show()