Skip to content

Increasing customer of time deposit campaign

Project Brief

You, as a data scientist in Personalized Marketing Team, are trying to find the opportunity to increase time deposit offering and you were assigning to analyze and develop presentation report for CEO.

Questions :

  • With the given dataset, how are you going to recommend to Business Unit (with compelling data-driven evidence) to boost the following metrics:
    • conversion rate = number of customers that purchases time deposit / total customers
    • sales volume = amount of time deposit volume
  • If you can collect more data, what would you collect to improve the recommendation or the metrics

Process

  • Setting Up
  • Load And Making EDA (Exploratory Data Analysis)
  • Cleaning and Preparing Data
  • Train, Test model
  • Select and Evaluate model
  • Conclusion

Setting Up

We will install the library for manipulate (tiduverse) and fro model training (caret xgboost randomForest)

library(tidyverse)
Hidden output 
install.packages("caret")
library(caret)
Hidden output 
install.packages("xgboost")
library(xgboost)
Hidden output 
install.packages("randomForest")
library(randomForest)
Hidden output

Load And Making EDA (Exploratory Data Analysis)

  • we will load dataset from .csv to dataframe.
  • we will split data to Categorical (feature) dataset and Numerical (feature) dataset and explore it.
df <- read_delim("datascience-full.csv", delim =  ";")
head(df)
summary(df)
## the dataset has no missing value
glimpse(df)
df %>%
	count(y)

Take Away

  • the dataset has 17 features (1 label) and 45,211 observations
  • 17 features including 10 categorical (charector) and 7 numerical (double)
  • the dataset has no missing value

Implementation note

  1. we will covert categorical to factor in cleaning process

Explore Categorical Features

‌
‌
‌