Problem Statement:
- Are you aware of what, when, and why your customers will make a purchase? Many businesses undertake an intense pursuit to discover these answers, dedicating valuable resources to data-driven campaigns and high-cost strategies – yet the actual outcomes often remain elusive and disappointing.
- Customer information is considered to be a valuable asset, however its true worth can only be established when it is used. Many companies have large collections of data that appear to be impressive, but upon further examination, they may contain outdated or unimportant information.
- Propensity modeling is a method that aims to forecast the chance that individuals, leads, and customers will engage in specific actions. This method uses statistical analysis which takes into account all the independent and confounding factors that impact customer behavior.
- Suppose you are working for a company as a Data Scientist. Your company is commissioned by an insurance company to develop a tool to optimize their marketing efforts.
This project is aimed at building a propensity model to identify potential customers.
Data:
The insurance company has provided you with a historical data set (train.csv). The company has also provided you with a list of potential customers to whom to market (test.csv). From this list of potential customers, you need to determine yes/no whether you wish to market to them. (Note: Ignore any additional columns available other than the listed below in the table)
Your focus in this project should be on the following:
The following is recommendation of the steps that should be employed towards attempting to solve this problem statement:
- Exploratory Data Analysis: Analyze and understand the data to identify patterns, relationships, and trends in the data by using Descriptive Statistics and Visualizations.
- Data Cleaning: This might include standardization, handling the missing values and outliers in the data.
- Dealing with Imbalanced data: This data set is highly imbalanced. The data should be balanced using the appropriate methods before moving onto model building.
- Feature Engineering: Create new features or transform the existing features for better performance of the ML Models.
- Model Selection: Choose the most appropriate model that can be used for this project.
- Model Training: Split the data into train & test sets and use the train set to estimate the best model parameters.
- Model Validation: Evaluate the performance of the model on data that was not used during the training process. The goal is to estimate the model’s ability to generalize to new, unseen data and to identify any issues with the model, such as overfitting.
- Model Deployment: Model deployment is the process of making a trained machine learning model available for use in a production environment.
Tasks/Activities List
Your code should contain the following activities/Analysis:
- Collect the time series data from the CSV file linked here.
- Exploratory Data Analysis (EDA) – Show the Data quality check, treat the missing values, outliers etc if any.
- Get the correct datatype for date.
- Balancing the data.
- Feature Engineering and feature selection.
- Train/Test Split – Apply a sampling distribution to find the best split.
- Choose the metrics for the model evaluation
- Try multiple classification models and choose the best one.
- Model Selection, Training, Predicting and Assessment
- Hyperparameter Tuning/Model Improvement
- Please add a column to the testingCandidate.csv file. In this column, for each observation indicate a 1 (yes) or a 0 (no) whether you wish to market to that candidate.
- Model deployment plan.
Success Metrics
Below are the metrics for the successful submission of this case study.
- The accuracy of the model on the test data set should be > 85% (Subjective in nature)
- Add methods for Hyperparameter tuning.
- Perform model validation.
Bonus Points
- You can package your solution in a zip file included with a README that explains the installation and execution of the end-to-end pipeline.
- You can demonstrate your documentation skills by describing how it benefits our company.