compare-models-py

Python

#Load library and data

import pandas as pd

from ChannelAttributionPro import *

import matplotlib.pyplot as plt

df_paths = pd.read_csv("https://app.channelattribution.io/data/Data.csv",sep=";")

#Set your token

token="yourtoken"

#Compare MTA models
 
res=compare_models(df_paths=df_paths,var_path="path",var_conv="total_conversions",
    var_null="total_null",cha_sep=">",perc_reall=0.1,perc_sample=0.10,
    max_nsim=10000,min_perc_traffic=0.005,niter=10,flg_extra_path=1,
    max_markov_order=4,seed=1234567,verbose=1)

res["performance"].columns=res["performance"].columns.str.replace('markov_model_order','markov')
plt.figure(figsize=(15, 6))
res["performance"].boxplot()
plt.savefig('performance_1.png', format='png')


#Compare MTA models and Reward model (UAM)

df_paths = pd.read_csv("https://app.channelattribution.io/data/data_paths_2.csv",sep=";")

df_ctr = pd.read_csv("https://app.channelattribution.io/data/data_ctr_2.csv",sep=";")

res=compare_models(df_paths=df_paths,df_ctr=df_ctr,channel_conv_name="((CONV))",perc_reall=0.1,perc_sample=0.10,
                   max_nsim=10000,min_perc_traffic=0.005,niter=10,flg_extra_path=1,max_markov_order=4,seed=1234567,verbose=1)

res["performance"].columns=res["performance"].columns.str.replace('markov_model_order','markov')
plt.figure(figsize=(15, 6))
res["performance"].boxplot()
plt.savefig('performance_2.png', format='png')