Practical Examination & Project Work50-Mark Rubric, 15 Lab Programs, SDG Capstone & Viva Voce
The definitive master guide for CBSE Class 10 AI Practical Examination (50 Marks). Access the official scoring rubric, the 15-program lab logbook checklist, three complete UN SDG capstone project templates, and 10 external examiner Viva Voce questions with model answers.
15 Marks
Python Coding Exam15 Marks
15 Verified Programs10 Marks
Portfolio & Model10 Marks
Lab + Project Viva1. Official CBSE 50-Mark Practical Examination Rubric
The external and internal examiners evaluate candidates strictly according to the five prescribed components below:
| Assessment Component | Description & Deliverables | Marks |
|---|---|---|
| 1. Hands-on Practical Examination | Writing and running Python programs in Jupyter Notebook (NumPy, Pandas, Matplotlib, OpenCV) and computer vision/NLP exercises. | 15 |
| 2. Practical File / Student Logbook | Physical journal containing minimum 15 teacher-verified Python programs with aim, source code, comments, and printout of sample outputs. | 15 |
| 3. Practical Viva Voce | Oral questioning by external examiner on Python commands, OpenCV functions, data handling, and confusion matrix formulas. | 5 |
| 4. Capstone AI Project / Portfolio | Documentation of an AI project aligned with UN Sustainable Development Goals (4Ws Canvas, dataset, model training, evaluation metrics). | 10 |
| 5. Project Viva Voce | Defending the student's specific project: problem statement, ethical considerations, data bias mitigation, and future improvements. | 5 |
| Total Practical Marks | Internal + External Board Assessment | 50 |
2. Required 15 Practical Programs Checklist
Verify that your practical journal includes every program listed below before external board submission. Full source code for each program is available in the Advance Python Chapter.
NumPy 1D and 2D Array Creation with shape, ndim, size attributes
NumPyArray Slicing, Reshaping (arange to 3x4 matrix), and Vector Arithmetic
NumPyStatistical Measures: Mean, Median, Variance, Standard Deviation, Percentiles
NumPyBuilding a Student Performance DataFrame and inspect head(3) and describe()
PandasDataFrame Conditional Filtering (Score >= 90) and Computed Columns
PandasHandling Missing Values: Detecting isna() and Imputing with mean fillna()
PandasLine Graph: Model Training Loss vs Validation Loss over 5 Epochs
MatplotlibBar Chart: Comparing Student Capstone Counts across 4 AI Domains
MatplotlibScatter Plot: Analyzing Correlation between Study Hours and AI Exam Scores
MatplotlibOpenCV: Image Reading with imread, Grayscale Conversion, and Resizing
OpenCVOpenCV: Applying Gaussian Blur and Canny Edge Detection Algorithm
OpenCVScikit-Learn: Confusion Matrix (TN, FP, FN, TP) and Metric Calculations
Scikit-LearnText Tokenisation and Stopword Removal using NLTK
NLTKPorter Stemmer vs WordNet Lemmatizer comparison on word lists
NLTKTF-IDF Vectorizer generation on a multi-sentence corpus
Scikit-Learn3. UN Sustainable Development Goals (SDG) Project Blueprints
CBSE mandates that student capstone projects solve a tangible challenge aligned with the United Nations Sustainable Development Goals. Here are three high-scoring project blueprints ready for your portfolio:
Project 1: Automated Coral Reef Bleaching Image Classifier
• Who: Marine biologists and reef conservation agencies.
• What: Severe coral bleaching due to rising sea surface temperatures.
• Where: Coral reef ecosystems (e.g. Great Barrier Reef, Lakshadweep).
• Why: Protect 25% of all marine biodiversity dependent on healthy coral reefs.
• Data: 600 underwater images categorized into Healthy, Bleached, and Dead.
• Model: MobileNet transfer learning via Teachable Machine.
• Metrics: 94.2% Accuracy, 91.5% Precision, 93.0% Recall.
• Ethics: Eliminates human diver fatigue while safeguarding data privacy.
Project 2: Urban Air Quality Index (AQI) & Smog Predictor
• Who: Urban citizens, asthmatic children, city pollution control boards.
• What: Deadly winter smog spikes exceeding PM2.5 safe limits by 15x.
• Where: Major metropolitan centers (Delhi NCR, Mumbai, Kanpur).
• Why: Enable proactive school closures and emergency traffic rationing.
• Data: CPCB continuous ambient air quality dataset (PM2.5, NO2, Wind Speed).
• Model: Linear Regression / Decision Tree in Python Pandas.
• Metrics: $R^2$ Score of 0.88, Mean Absolute Error (MAE) = 14.2 AQI points.
• Ethics: Transparent open-source data prevents municipal under-reporting bias.
Project 3: Cyberbullying & Toxic Sentiment Comment Filter
• Who: School students, teenagers, and educational online forums.
• What: Severe psychological distress caused by hateful online harassment.
• Where: School virtual learning portals and social platforms.
• Why: Foster a safe, inclusive digital learning space for all learners.
• Data: 2,000 labelled user comments cleaned with NLTK stopwords.
• Model: TF-IDF feature extraction with Multinomial Naive Bayes.
• Metrics: Precision = 92.4% (minimizing wrongful censorship of benign text).
• Ethics: Strict anonymization of user handles to protect student identity.
4. External Examiner Viva Voce Master Preparation
Review these official model answers to questions most frequently posed by external examiners during the Class 10 practical board viva.
Frequently Asked Questions (FAQ)
Get Your Capstone AI Project & Lab File Verified by Experts
Book a 1-on-1 personalized mentorship session to get your UN SDG project audited, review your 15-program lab file, and conduct full mock viva voce interviews before your final board exam.