Forum Discussion
getting challenges with data science
First, don't be too hard on yourself. Scholarship selections are often based on many factors beyond a single project, and without seeing the reviewers' feedback, it's impossible to know exactly why you weren't selected.
From what you've described, you've already covered many of the core technical steps of a Data Science project: data cleaning, preprocessing, exploratory analysis, feature engineering, model training, and evaluation. What often distinguishes stronger projects is demonstrating business impact, clearly explaining why certain features and models were chosen, comparing multiple algorithms, addressing class imbalance, performing robust validation, and presenting actionable insights rather than just model metrics.
To gain more practical experience, I'd recommend building additional end-to-end projects, participating in Kaggle competitions, creating dashboards to communicate results, and sharing your work on GitHub with detailed documentation. Focus on solving real-world problems and explaining your thought process, not just the code. Employers and scholarship reviewers often value problem-solving, storytelling, and reproducibility as much as technical implementation.
Keep learning and building, one scholarship outcome doesn't define your potential in Data Science or Machine Learning. Every project you complete strengthens your skills and portfolio, and consistency over time is what ultimately makes the biggest difference. 🚀
Thanks for sharing your experience. First of all, completing an end-to-end Customer Churn Prediction project is a solid achievement—it demonstrates initiative and covers many of the core stages of a data science workflow.
That said, scholarship and project evaluations often look beyond whether a model was built. Some areas that commonly make a project stand out include:
Clearly defining the business problem and explaining how the model creates value.
Demonstrating strong feature engineering with well-reasoned feature selection.
Comparing multiple baseline and advanced models, along with proper hyperparameter tuning.
Addressing class imbalance, data leakage, and overfitting.
Explaining model decisions using interpretability techniques such as SHAP or feature importance.
Validating the model with robust evaluation methods (cross-validation, appropriate metrics, error analysis).
Showing how the solution could be deployed, monitored, and maintained in a real production environment.
Presenting results through a clean dashboard or report that communicates insights to non-technical stakeholders.
One suggestion is to ask the scholarship organizers if they can provide feedback on your submission. Even brief comments can help identify the specific areas to improve.
I'd also like to hear from the community: What skills or project elements have made the biggest difference in your own scholarship applications or entry-level Data Science interviews? Any recommendations for gaining more real-world, hands-on experience would be greatly appreciated.