import streamlit as st import pandas as pd import numpy as np import datetime as dt from dateutil.relativedelta import relativedelta import pandas as pd import plotly.express as px # pip install plotly-express import streamlit as st # from pyspark import SparkContext, SparkConf # from pyspark.sql import SparkSession # from pyspark.sql.types import StructType # from pyspark.sql.types import StructField # from pyspark.sql.types import StringType import matplotlib.pyplot as plt #import seaborn as sns import altair as alt from functions import * from datetime import datetime, timedelta, date import plotly.figure_factory as ff # emojis: https://www.webfx.com/tools/emoji-cheat-sheet/ st.set_page_config(page_title="Media Planing", layout="wide",page_icon="📊") #page_icon=":bar_chart:", #st.set_page_config(page_title="Mapping Demo", page_icon="🌍") st.title('"Media Planning"') option=st.sidebar.selectbox( 'some text here', ('first','one day', 'couple of days', "some specefic ad/ads")) st.sidebar.write('You selected:', option) df=pd.read_csv('pages/ads_historic.csv') # st.dataframe(df) if option=="first": pass if option=="one day": d = st.sidebar.date_input( "Enter that day", datetime(2022, 10, 6)) st.sidebar.markdown("""---""") one_day=df[df['day']==str(d)][["add_name" ,"day"]] # cercle one_day=one_day.groupby(pd.Grouper(key="add_name")).count().reset_index() one_day.rename(columns={'day': 'frequency'}, inplace=True) labels = one_day['add_name'].tolist() sizes = one_day['frequency'].tolist() #age = st.sidebar.slider('choose how many ads are the most accured?', 0, 10, 2) if len(sizes)==0: #one_day.shape == (0,0): pass else: age = st.sidebar.slider('choose how many ads are the most accured?', 0, 10, 2) kpi1, kpi2 = st.columns(2) with kpi1: st.write('the Ads that occur at least ',age,' times are :') st.json(get_most_ads(age,one_day)) one_day=df[df['day']==str(d)][["add_name" ,"day"]] one_day=one_day.groupby(pd.Grouper(key="add_name")).count().reset_index() one_day.rename(columns={'day': 'frequency'}, inplace=True) with kpi2: kpi2.metric( label="Number of Ads ⏳", value=round(len(one_day['add_name'].tolist()))) st.markdown("""---""") kp1, kp2 = st.columns(2) with kp1: st.write('The partition of the all Ads on ',d) fig1, ax1 = plt.subplots() ax1.pie(sizes, labels=labels, autopct='%1.1f%%', shadow=True, startangle=90) ax1.axis('equal') st.pyplot(fig1) with kp2: chart = ( alt.Chart(one_day) .mark_bar() .encode( x=alt.X("frequency", type="quantitative", title="frequency"), y=alt.Y("add_name", type="nominal", title="Ads name"), ).properties(title='Sales of consumer goods') ) st.altair_chart(chart, use_container_width=True) if option=="couple of days": kpi1, kpi2 = st.columns(2) start = st.sidebar.date_input(label='Start: ', value=datetime(year=2021, month=5, day=20, hour=16, minute=30), key='#start', help="The start date time", on_change=lambda : None) end = st.sidebar.date_input(label='End: ', value=datetime(year=2021, month=5, day=30, hour=16, minute=30), key='#end', help="The end date time", on_change=lambda : None) st.sidebar.write('Start: ', start, "End: ", end) st.sidebar.markdown("""---""") #age = st.sidebar.slider('choose how many ads are the most accured?', 0, 10, 2) df=pd.read_csv('pages/ads_historic.csv') df['day']=df['day'].apply(convert_date) df=df[(df['day']>=datetime.strptime(str(start), '%Y-%M-%d')) & (df['day']<=datetime.strptime(str(end), '%Y-%M-%d'))] df=pd.read_csv('pages/ads_historic.csv') df['day']=df['day'].apply(convert_date) df=df[(df['day']>=datetime.strptime(str(start), '%Y-%M-%d')) & (df['day']<=datetime.strptime(str(end), '%Y-%M-%d'))] one_day=df[["add_name" ,"day"]] # cercle one_day=one_day.groupby(pd.Grouper(key="add_name")).count().reset_index() labels = one_day['add_name'].tolist() one_day.rename(columns={'day': 'frequency'}, inplace=True) if len(labels)==0: pass else: age = st.sidebar.slider('choose how many ads are the most accured?', 0, 10, 2) with kpi1: st.write('the Ads that occur at least ',age) st.json(get_most_ads(age,one_day)) with kpi2: kpi2.metric( label="Number of Ads ⏳", value=round(len(labels)) ) sizes = one_day['frequency'].tolist() st.markdown("""---""") kp1, kp2 = st.columns(2) with kp1: fig1, ax1 = plt.subplots() ax1.pie(sizes, labels=labels, autopct='%1.1f%%', shadow=True, startangle=90) ax1.axis('equal') st.pyplot(fig1) with kp2: chart = ( alt.Chart(one_day) .mark_bar() .encode( x=alt.X("frequency", type="quantitative", title="frequency"), y=alt.Y("add_name", type="nominal", title="Ads name") ) ) st.altair_chart(chart, use_container_width=True) df=pd.read_csv('pages/ads_historic.csv') result = pd.DataFrame() options2=df['add_name'].unique().tolist() for pub in options2: result=pd.concat([result,df[df['add_name']==pub]]) gkk = result.groupby(['add_name','day']).size().reset_index() gkk.rename(columns={'day': 'frequency'}, inplace=True) gkk.rename(columns={0: 'value'}, inplace=True) chart = alt.Chart(gkk).mark_line().encode( x=alt.X('frequency:N'), y=alt.Y('value:Q'), color=alt.Color("add_name:N") ).properties(title="Hello World") st.altair_chart(chart, use_container_width=True) if option=="some specefic ad/ads": option1 = st.sidebar.selectbox( 'How many ads do you want to see', ('One', 'Many')) if option1=="One": df=pd.read_csv('pages/ads_historic.csv') ads_name=df['add_name'].unique().tolist() genre = st.sidebar.radio( "Select the Ad", ads_name) #df=pd.read_csv('pages/ads_historic.csv') ad_name=df[df['add_name']==genre] each_day=ad_name.groupby(pd.Grouper(key="day")).count().reset_index()[['day','add_name']] st.write('the max value is ',max_min(each_day)[0],' on ',max_min(each_day)[1],' and the min value is ',max_min(each_day)[2],' on ',max_min(each_day)[3]) each_day.rename(columns={'add_name': 'frequency'}, inplace=True) st.line_chart(each_day, x="day", y="frequency", width=0, height=0, use_container_width=True) if option1=="Many": df=pd.read_csv('pages/ads_historic.csv') ads_name=df['add_name'].unique().tolist() st.sidebar.markdown("""---""") options2 = st.sidebar.multiselect('Choose the Ads',ads_name) result = pd.DataFrame() for pub in options2: result=pd.concat([result,df[df['add_name']==pub]]) gkk=pd.DataFrame() gkk1=pd.DataFrame() #if result.shape != (0,0): if result.shape == (0,0): pass else: gkk = result.groupby(['add_name','day']).size().reset_index() gkk.rename(columns={0: 'frequency'}, inplace=True) gkk1 = result.groupby(['add_name','day']).size().reset_index() gkk1.rename(columns={0: 'frequency'}, inplace=True) #st.dataframe(data=gkk,use_container_width=False) k1 ,k2=st.columns(2) with k1: gkk=gkk.groupby(pd.Grouper(key="add_name")).sum().reset_index() sizes=gkk['frequency'].tolist() labels=gkk['add_name'].tolist() #st.write('The partition of the all Ads on ',d) fig1, ax1 = plt.subplots() ax1.pie(sizes, labels=labels, autopct='%1.1f%%', shadow=True, startangle=90) ax1.axis('equal') st.pyplot(fig1) with k2: gkk=gkk.groupby(pd.Grouper(key="add_name")).sum().reset_index() chart = ( alt.Chart(gkk) .mark_bar() .encode( x=alt.X("frequency", type="quantitative", title="frequency"), y=alt.Y("add_name", type="nominal", title="Ads name"), ).properties(title='Sales of consumer goods') ) st.altair_chart(chart, use_container_width=True) st.markdown("""---""") #st.dataframe(data=gkk1,use_container_width=False) chart = alt.Chart(gkk1).mark_line().encode( x=alt.X('day:N'), y=alt.Y('frequency:Q'), color=alt.Color("add_name:N") ).properties(title="Hello World") st.altair_chart(chart, use_container_width=True) # df=pd.read_csv('pages/ads_historic.csv') # ads_name=df['add_name'].unique().tolist() # options2 = st.sidebar.multiselect('Choose the Ads',ads_name) # result = pd.DataFrame() # df=pd.read_csv('pages/ads_historic.csv') # for pub in options2: # #print(pub) # result=pd.concat([result,df[df['add_name']==pub]]) # #result.append(df[df['add_name']==pub]) # #st.dataframe(data=result,use_container_width=False) # gkk = result.groupby(['add_name','day']).size().reset_index() # gkk.rename(columns={'day': 'frequency'}, inplace=True) # #st.dataframe(data=gkk,use_container_width=False) # labels=gkk['add_name'].unique().tolist() # hist_data =[] # for la in labels: # hist_data.append(np.array(gkk[gkk['add_name']==la][0].tolist())) # group_labels = labels # # Create distplot with custom bin_size # fig = ff.create_distplot( # hist_data, group_labels) # # Plot! # st.plotly_chart(fig, use_container_width=True)