Inspiration
The inspiration behind our project for the NUS Datathon sponsored by Singlife stems from the observation of a concerning trend in the insurance industry. Singlife noticed that potential policyholders often hesitate and eventually disengage during the insurance acquisition process. This raised the need to leverage Singlife's dataset to derive actionable insights that could enhance the overall customer experience.
What it does
The current project objective is to deconstruct the Singlife dataset in order to comprehensively identify important points of contact that lead to customer churn. Therefore, we searched for opportunities that could facilitate the general application process and deliver individual customer satisfaction forecasting as well as conversion rates. The objective is to make Singlife’s market position stronger by eliminating pain points in the customer journey.
How we built it
To tackle the challenge, we utilized a combination of data analysis, machine learning, and data visualization techniques. We explored and cleaned the dataset, identified relevant features, and implemented predictive models to forecast customer satisfaction and conversion rates. The project involved collaborative efforts in coding, testing, and refining the models.
Challenges we ran into
One of the primary challenges we faced was navigating the complexity of the dataset and understanding the intricacies of customer behavior. Cleaning and preprocessing the data to ensure its suitability for modeling posed another significant challenge. Additionally, fine-tuning the machine learning models to achieve optimal predictive performance required iterative adjustments.
Accomplishments that we're proud of
We have successfully developed 5 predictive models(Logistic Regression, KNN, Decision Tree, Random Forest, and Neural Network) that shed light on critical touchpoints affecting customer satisfaction and conversion rates. One algorithm that stands out as a source of pride is the Random Forest, boasting an impressive accuracy rate of 88%. Our team's collaborative effort in overcoming challenges and achieving a meaningful outcome is an accomplishment in itself.
What we learned
Throughout the datathon, we gained valuable insights into the insurance industry and the importance of leveraging data to address customer experience challenges. We enhanced our skills in data analysis, machine learning, and teamwork through real-world problems. The experience also reinforced the significance of a customer-centric approach to business solutions.
What's next
Looking ahead, we plan to further refine our predictive model based on feedback received during the datathon. Additionally, we aim to explore more advanced machine learning techniques and implement additional features to enhance the accuracy and interpretability of our predictions.
Thanks Singlife and SDS for organizing this insightful and valuable datathon!
Log in or sign up for Devpost to join the conversation.