Data Science & Machine Learning

What you will be able to do

• Write clean, reusable Python for data work with NumPy and pandas
• Explore and visualise data to find patterns, problems and questions worth modelling
• Apply descriptive and inferential statistics, including confidence intervals and hypothesis tests
• Engineer features and build leakage-safe preprocessing pipelines
• Train, tune and compare regression, classification and clustering models with scikit-learn
• Choose evaluation metrics that match the problem and validate models honestly
• Assess fairness, privacy and explainability, and document a model with a model card
• Deliver an end-to-end capstone project and explain its results and limitations to a non-technical audience

About this course

Syllabus

This syllabus is a draft awaiting instructor review. Topics, order and schedule may change before the first batch is announced.

1. 01Python for data science+4 live sessions · 8 hours · A fast Python refresher for people who already program, plus NumPy and a reproducible project setup.

   Learning objectives

   • Set up Python, virtual environments, Jupyter and Git for reproducible work
   • Use Python’s core types, functions and modules idiomatically
   • Work with NumPy arrays and vectorised operations
   • Decide what belongs in a notebook and what belongs in a reusable module

   Live sessions

   1. 01Setup and Python core types2 hInstalling Python, venv and pip, Jupyter and VS Code, Git basics; numbers, strings, lists, tuples, dictionaries and sets; control flow.
   2. 02Functions, modules and clean code2 hFunctions and arguments, comprehensions, modules and imports, basic type hints, reading tracebacks, moving notebook code into reusable functions.
   3. 03NumPy and vectorised computation2 hArrays, shapes and data types, indexing and slicing, broadcasting, vectorised maths and aggregations, random number generation with seeds for reproducibility.
   4. 04Python review and doubt-clearing2 hSolving small data-processing problems with plain Python and NumPy; open questions.

   Practice

   Write a small module of tested helper functions that summarise a numeric dataset with NumPy.

   Assessment

   Quiz on Python types, functions and NumPy, plus the helper module as an assignment.

   You will be able to

   You can write clean, reproducible Python and NumPy code for data work.

2. 02Data wrangling with pandas+5 live sessions · 10 hours · Load, clean, combine and reshape real-world data with pandas, and handle it responsibly.

   Learning objectives

   • Load data from CSV, Excel, SQL databases and web APIs
   • Select, filter and transform data with correct data types
   • Handle missing values, duplicates, outliers, text and dates
   • Group, merge and reshape data, including simple time series
   • Check a dataset’s licence and minimise personal data before analysis

   Live sessions

   1. 05Loading data and data provenance2 hReading CSV, Excel and JSON files, querying a SQL database from pandas, calling a public API; dataset licences and terms of use; identifying and minimising personal data.
   2. 06Selecting, filtering and transforming2 hDataFrames and Series, loc and iloc, boolean filtering, data types and conversion, new columns with vectorised operations, method chaining.
   3. 07Cleaning messy data2 hMissing values and when to drop or impute, duplicates, inconsistent categories, outliers, string methods, parsing dates, validating assumptions with assertions.
   4. 08Grouping, merging, reshaping and time series2 hgroupby and aggregation, merge and join types, pivot_table and melt, resampling and rolling windows for time series.
   5. 09pandas review and doubt-clearing2 hCleaning a deliberately messy public dataset end to end and documenting every step; open questions.

   Practice

   Clean and combine two related public datasets into one analysis-ready table with a written cleaning log.

   Assessment

   Assignment: the cleaning notebook and log, reviewed for correctness and reproducibility.

   You will be able to

   You can turn raw, messy data into a clean, documented dataset with pandas.

3. 03Exploratory analysis and visualisation+3 live sessions · 6 hours · Explore data systematically and communicate what you find with clear, honest charts.

   Learning objectives

   • Follow a structured exploratory analysis workflow
   • Choose the right chart for distributions, comparisons and relationships
   • Build static and interactive charts that are accessible and not misleading

   Live sessions

   1. 10Exploratory data analysis with matplotlib and seaborn2 hUnivariate and bivariate exploration, histograms, box plots, scatter plots and heatmaps, spotting data problems and forming questions.
   2. 11Honest and interactive visualisation2 hChoosing chart types, axes and scales that do not mislead, accessible colour and labelling, interactive charts with Plotly.
   3. 12EDA workshop and doubt-clearing2 hA timed exploratory analysis on a new dataset presented in five minutes; peer feedback; open questions.

   Practice

   Produce an exploratory report on a public dataset with at least six labelled charts and the questions they raise.

   Assessment

   Assignment: the exploratory report, reviewed for insight and chart honesty.

   You will be able to

   You can explore an unfamiliar dataset and communicate what it shows with clear charts.

4. 04Statistics for data science+6 live sessions · 12 hours · The probability and statistics needed to reason about uncertainty, test claims and avoid common traps.

   Learning objectives

   • Describe data with measures of centre, spread and shape
   • Explain probability distributions, sampling and the central limit theorem
   • Construct confidence intervals and run and interpret hypothesis tests
   • Design and analyse a simple A/B test and avoid common pitfalls
   • Distinguish correlation from causation and fit a simple linear regression

   Live sessions

   1. 13Descriptive statistics and distributions2 hMean, median, variance, standard deviation and percentiles; skew and outliers; normal, binomial and Poisson distributions with simulations in NumPy.
   2. 14Probability, sampling and the central limit theorem2 hProbability rules, conditional probability and Bayes’ rule, sampling and sampling bias, the central limit theorem demonstrated by simulation, bootstrapping.
   3. 15Confidence intervals and hypothesis testing2 hStandard error, confidence intervals, null and alternative hypotheses, t-tests and chi-square tests with SciPy, what a p-value does and does not mean.
   4. 16A/B testing and statistical pitfalls2 hDesigning an experiment, sample size and power, effect size versus significance, peeking and multiple comparisons, Simpson’s paradox.
   5. 17Correlation, causation and linear regression2 hCorrelation coefficients, confounding, simple and multiple linear regression with statsmodels, interpreting coefficients and residuals.
   6. 18Statistics review and doubt-clearing2 hWorking through applied problems on synthetic experiment data; open questions.

   Practice

   Analyse a synthetic A/B test dataset: check the design, run the test, report the effect with a confidence interval and state the limitations.

   Assessment

   Timed quiz on distributions, confidence intervals and hypothesis tests, plus the A/B test write-up as an assignment.

   You will be able to

   You can use statistics to quantify uncertainty and judge whether a claim holds up.

5. 05Machine learning foundations+4 live sessions · 8 hours · What machine learning is, how to set up a problem correctly and the core scikit-learn workflow.

   Learning objectives

   • Frame a problem as regression, classification or clustering, or decide that ML is not needed
   • Split data correctly and recognise data leakage
   • Train and interpret linear and logistic regression models with scikit-learn
   • Compare models against a simple baseline

   Live sessions

   1. 19Framing machine-learning problems2 hSupervised and unsupervised learning, targets and features, training, validation and test sets, baselines, data leakage and how it hides in real projects, when not to use ML.
   2. 20Linear regression and regularisation with scikit-learn2 hThe fit and predict interface, linear regression, overfitting and underfitting, ridge and lasso regularisation, interpreting coefficients.
   3. 21Logistic regression and classification2 hBinary and multi-class classification, logistic regression, predicted probabilities and decision thresholds, a first look at the confusion matrix.
   4. 22ML foundations review and doubt-clearing2 hFinding the leakage in a set of flawed example projects; reviewing learners’ problem framings; open questions.

   Practice

   Frame two problems from public datasets, build baselines and train linear and logistic regression models for them.

   Assessment

   Quiz on problem framing and leakage, plus a notebook comparing each model with its baseline.

   You will be able to

   You can set up a machine-learning problem correctly and train simple, interpretable models.

6. 06Feature engineering and pipelines+3 live sessions · 6 hours · Turn raw columns into useful features and wrap preprocessing in pipelines that cannot leak.

   Learning objectives

   • Encode categorical data, scale numeric data and impute missing values appropriately
   • Create features from dates, text and aggregations
   • Build leakage-safe preprocessing with Pipeline and ColumnTransformer

   Live sessions

   1. 23Encoding, scaling and imputation2 hOne-hot and ordinal encoding, high-cardinality categories, standardisation and normalisation, imputation strategies and missing-value indicators.
   2. 24Creating and selecting features2 hDate and time features, simple text features with bag-of-words, ratios and group aggregations, domain knowledge, feature selection and the curse of dimensionality.
   3. 25Pipelines, ColumnTransformer and doubt-clearing2 hComposing preprocessing and models with Pipeline and ColumnTransformer, fitting only on training data, saving and loading pipelines; open questions.

   Practice

   Rebuild an earlier model as a full pipeline with engineered features and measure the change in validation performance.

   Assessment

   Assignment: a pipeline notebook with a short note explaining why each feature should help.

   You will be able to

   You can engineer features and build preprocessing pipelines that are safe to reuse.

7. 07Model evaluation and tuning+3 live sessions · 6 hours · Measure models with the right metrics, validate them honestly and tune them without fooling yourself.

   Learning objectives

   • Choose regression and classification metrics that match the cost of errors
   • Use cross-validation and hyperparameter search correctly
   • Diagnose bias and variance, analyse errors and handle imbalanced data

   Live sessions

   1. 26Evaluation metrics2 hMAE, RMSE and R² for regression; accuracy, precision, recall, F1, ROC-AUC and precision-recall curves for classification; choosing thresholds from the cost of errors; calibration.
   2. 27Cross-validation and hyperparameter tuning2 hk-fold, stratified and time-series cross-validation, grid and randomised search, nested validation, learning curves and the bias–variance trade-off.
   3. 28Error analysis, imbalanced data and doubt-clearing2 hInspecting where a model fails, class imbalance and class weights, resampling and its risks, keeping the test set untouched until the end; open questions.

   Practice

   Evaluate and tune a classifier on an imbalanced public dataset, choose a threshold from stated costs and report results on the held-out test set.

   Assessment

   Timed quiz on metrics and validation, plus an evaluation report as an assignment.

   You will be able to

   You can evaluate and tune models honestly and explain what the numbers mean.

8. 08Machine-learning algorithms+5 live sessions · 10 hours · How the main families of algorithms work, when to use each, and how to find structure without labels.

   Learning objectives

   • Explain and apply decision trees, random forests and gradient boosting
   • Compare k-nearest neighbours, support vector machines and naive Bayes
   • Describe how a simple neural network learns and where deep learning fits
   • Apply clustering and dimensionality reduction to unlabelled data

   Live sessions

   1. 29Decision trees2 hHow trees split, impurity measures, depth and pruning, interpreting a tree, why single trees overfit.
   2. 30Ensembles: random forests and gradient boosting2 hBagging and random forests, boosting and histogram-based gradient boosting in scikit-learn, key hyperparameters, out-of-bag estimates, comparing ensembles with simpler models.
   3. 31Other supervised methods and an introduction to neural networks2 hk-nearest neighbours, support vector machines and naive Bayes compared; how a multi-layer perceptron learns with gradient descent; when deep learning is and is not worth it.
   4. 32Unsupervised learning: clustering and PCA2 hk-means and choosing k, hierarchical clustering, DBSCAN, principal component analysis for visualisation and compression, interpreting clusters carefully.
   5. 33Algorithms review and doubt-clearing2 hChoosing an algorithm for a set of scenario briefs; comparing learners’ model results; open questions.

   Practice

   Compare at least four algorithms on one supervised problem using the same pipeline and validation, and segment a synthetic customer dataset with clustering.

   Assessment

   Project milestone: a model comparison notebook with a justified recommendation, reviewed with feedback.

   You will be able to

   You can choose, train and compare suitable algorithms for supervised and unsupervised problems.

9. 09Responsible AI and communicating results+3 live sessions · 6 hours · Check models for fairness and privacy risks, explain their behaviour and communicate results honestly.

   Learning objectives

   • Identify sources of bias in data and models and measure performance across groups
   • Apply privacy principles, including data minimisation and the limits of anonymisation
   • Explain model behaviour with permutation importance, partial dependence and SHAP
   • Document a model with a model card and present results to non-technical audiences

   Live sessions

   1. 34Fairness, bias and privacy2 hWhere bias enters a project, proxy variables, comparing error rates across groups, fairness trade-offs; consent, purpose limitation and data minimisation under India’s DPDP Act 2023; re-identification risk; using AI assistants without sharing personal data.
   2. 35Explainability and model documentation2 hGlobal and local explanations, permutation importance, partial dependence plots, SHAP values and their limits, model cards and datasheets for datasets, human oversight of model decisions.
   3. 36Communicating and sharing results, with doubt-clearing2 hWriting for decision-makers, presenting uncertainty, a simple Streamlit demo that shows limitations, saving and versioning models; open questions.

   Practice

   Audit an earlier model for performance differences across groups, explain its predictions and write a model card for it.

   Assessment

   Assignment: a responsible-AI review and model card for one of your models, reviewed with feedback.

   You will be able to

   You can assess, explain and document a model responsibly and communicate its results honestly.

10. 10Capstone project and final assessment+4 live sessions · 8 hours · Plan, build and present an end-to-end data-science project, with structured reviews and the final assessment.

    Learning objectives

    • Turn a brief into a well-framed problem with a realistic scope
    • Apply the full workflow from data audit to responsible evaluation
    • Present and defend methods and results in a live review

    Live sessions

    1. 37Capstone kickoff and data audit2 hWalking through the briefs; framing the problem, success measure and potential harms; auditing the data source, licence and personal data; agreeing scope and milestones.
    2. 38Capstone build lab and mid-point review2 hGuided build time with individual reviews of exploratory analysis, pipelines and evaluation plans; agreeing what to fix before submission.
    3. 39Final assessment2 hWritten and practical assessment across the whole course (pass mark 60%).
    4. 40Capstone presentations and final review2 hLive presentations to a non-technical audience, method walkthroughs and feedback; next steps for continued learning.

    Practice

    Build the capstone with regular commits, a reproducible notebook and module structure and an up-to-date README.

    Assessment

    Capstone project submission and live presentation, and the final assessment (pass mark 60%).

    You will be able to

    You can independently deliver and explain an end-to-end, responsibly evaluated data-science project.

Teaching formats and fees

Every format teaches the full syllabus live, with the same projects, assessments and certificate rules. The fee is per student for the whole course.

Launch fee — GST treatment to be confirmed.

Fees are set in Indian rupees. Amounts in other currencies are estimates; the amount and currency you would pay are shown before payment once online enrolment opens.

Schedules and batches

Choose the schedule that fits your week. Both teach the same syllabus and the same 80 live hours; only the number of weeks differs.

• Weekday Intensive classes are one hour, so each 2 h syllabus session is taught across consecutive weekday classes.
• Weekend Evening classes run on Saturday and Sunday evenings; the exact time and time zone are set before a batch is confirmed.
• Public holidays are not class days unless the instructor confirms a class. Holidays, instructor leave and cancelled classes move the end date later so no live hours are lost.
• Plan about 6 hours a week of exercises, assignments and project work outside class.
• Group, Duo and One-to-One courses all follow one of these two schedules.

Final project

Certificate

Tools and skills

Where these skills apply

• Data analysis and statistical reporting
• Building and evaluating predictive models
• Experiment design and A/B test analysis
• Customer, operations and risk analytics
• Preparing data and features for machine-learning systems
• Reviewing models for fairness, privacy and explainability

We help you prepare for work; we do not guarantee jobs, placements or salaries. Career resources

Questions about Data Science & Machine Learning