CB0494 Introduction to Data Science and AI Assignment Questions | NTU

Assignment Brief

1. Objectives

• To relate potential real-life problems with data science.
• To apply what has been taught and learnt in CB0494 on a real dataset and perform data analysis. The workflow is very important. This includes why your team uses certain tools and how these tools can help in your analysis.
• To conclude potential findings with predictions as part of the solution for the problem.

This mini project is to demonstrate your understanding of the tutorials and the data science content from your lecture slides and to be competent in basic data science analysis. The focus is still on the thought process and each team’s creativity in bringing the best out of the dataset with the tools taught in the course for the problem statement indicated by the team.

2. Assessment Criteria

(Reminder: Overall weightage is 30%.)

(a) Project Content and Analysis:

(b) Project Report:

(c) Presentation:

Dataset 2: Regarding data on lung diseases (From Kaggle):

URL: https://www.kaggle.com/datasets/samikshadalvi/lungs-diseases-dataset

Dataset: lung_disease_data.csv

Data Description: Please refer to the URL for information on the data fields, etc

Upon selecting your dataset, explain what real problem your team is going to solve and relate it to the data science questions. Study the dataset and select the appropriate data for your analysis via the Jupyter notebook. Each team should select the suitable tools taught in class to analyse the data and make sense of the analysis related to the problem statement they have stated, and eventually conclude their findings with some prediction.

To analyse the data, it should follow the thought process as taught in class. This means that each team need to strive to provide proof and justification as to why certain predictors are selected, for example, and obviously, each team should select predictors to have the best goodness of fit score as possible.

Each team does not need to be so “obsessed” about getting the goodness of fit score to an exceptionally high value when it may be impossible for some cases due to the quality of the dataset, which is beyond the control of the team. However, the thought process on how each team try the best possible way to analyse the data with the tools taught and, hence obtain a justified goodness of fit score is still important.

Each team can identify the limitations of the work and list out some recommendations for future work. The recommendations may include suggestions of using some tools (not taught in class) with explanations.

Please do not blindly follow through the steps without justification and reasons. That will certainly mean a F grade as it does not demonstrate any understanding of the class and hence failed in the application of the tools taught for the data analysis. Incorrect use of tools or using the tools in an unsuitable way will have marks deducted.

Each team must first convince themselves that the thought process and result of their mini project are realistic, reasonable and meaningful before they can convince me about that.

Please organise your programming codes and make suitable comments in your mini project Jupyter notebook file before submission so that it can be easily understood and readable. There is no limit in the length of your Jupyter notebook file for your mini project.

Problem Statement:

Lung diseases remain a significant public health concern in Singapore, particularly among the elderly and chronically ill populations. While access to healthcare is robust, patient recovery outcomes still vary due to multiple clinical and demographic factors. This project aims to develop a predictive model to classify whether a patient is likely to recover from lung disease based on their medical and demographic information. By identifying patients at risk of non-recovery early, healthcare providers can tailor interventions, reduce complications, and improve overall treatment efficiency.