Data Science With Machine Learning & AI Career Accelerator

What is statistical thinking? Why does it matter? What value does this approach add to business?

The best data scientists critically engage with the business problem before working on their analysis. This course will teach you how to identify potential issues at the analysis and solution planning stage that could result in huge costs to the business or have wider ethical implications.

You will understand how to select and apply relevant statistical techniques to create the simplest solutions to a range of business problems, saving the business time and money.

You will then continue to look at feature engineering and consider how it can be used to add value before diving into unsupervised learning techniques and how they can be used to solve business problems.

Course 1 learning outcomes

By the end of this course you will be able to:

• Systematically analyse complex organisational challenges to enhance problem-solving skills and propose strategic solutions
• Develop advanced critical and statistical thinking skills by identifying meaningful patterns in complex data
• Evaluate and select statistical techniques appropriate to a solutions design approach
• Synthesise and systematically implement techniques to enhance data representation and improve model performance
• Critically select, evaluate, and implement effective unsupervised learning techniques to design innovative solutions.

Week 1: Developing commercial awareness through statistical thinking

This week explores commercial thinking, collaboration, communication, critical thinking, problem-solving, and project management skills, then lays the foundation for selecting and applying statistical analysis techniques.

By framing statistical thinking and feature engineering as tools for developing human problem-solving and critical thinking skills, you will begin to make connections between theoretical concepts – statistics and machine learning (ML) techniques – and practical applications in business.

Week 2: Applying advanced statistical techniques for data science

This week explores hypothesis testing using Python, as well as how to find causal relationships and make predictions with regression techniques.

Week 3: Taking a critical approach to selecting statistical techniques

This week explores alternative methodologies and highlights the importance of critical selection of appropriate techniques.

Week 4: Engineering features and reducing dimensions

This week provides hands-on experience in feature engineering techniques, including developing the ability to extract meaningful features from real-life organisational data to address specific challenges. The latter part of the week focuses on transforming high-dimensional data into visual stories using powerful techniques like PCA and t-SNE.

Week 5: Detecting anomalies with unsupervised learning

This week, you will practise anomaly detection, apply decision-making skills to real-world scenarios, perform fraud detection, and communicate insights that drive organisational success.

Week 6: Performing clustering with unsupervised learning

This week focuses on performing clustering using unsupervised learning techniques. The two main clustering techniques you will examine include k-means clustering and hierarchical clustering. The content covers the concepts, methods, and practical implementation of both techniques, along with considerations for cluster evaluation and real-world applications.

Example Project

Customer segmentation using clustering techniques

Total duration: 12 hours

What? Customer segmentation enables businesses to understand their customers better, tailor their strategies to specific segments, and ultimately drive growth, profitability, and customer satisfaction.

Why? Select, test, and apply effective unsupervised learning techniques to solve business problems.

How? Segregate customers based on their demographic, behavioural, and transactional data into distinct clusters to improve marketing strategies and enhance business profitability, using clustering techniques in Python.

This comprehensive supervised learning course offers a strategic blend of ML modelling and practical business applications, equipping you to unravel valuable insights from complex data sets.

The focus is on an array of regression, classification, and boosting algorithms, as well as sophisticated ensemble techniques. You will practise translating data insights into actionable decisions that drive business excellence across a spectrum of organisational contexts. Through the exploration of deep learning techniques, you will discover how to draw hidden insights from intricate data structures.

Practical activities include crafting and fine-tuning advanced ML models tailored to diverse organisational scenarios.

Actionable results will be achieved through data preprocessing skills, strategic model selection, and insightful analysis, to lead to proposed solutions and the formulation of comprehensive strategies that resonate with business objectives.

Course 2 learning outcomes

By the end of this course you will be able to:

• Enhance problem-solving skills through analysing diverse organisational scenarios and accurately identifying key issues
• Inform data-driven decision-making with regression, classification, boosting algorithms, and ensemble techniques
• Develop accurate predictions across diverse organisational scenarios by building and testing advanced ML models
• Leverage deep learning across a range of data types to extract valuable insights from complex data
• Propose and present effective solutions to organisational problems using data preprocessing, model selection, and insightful analysis techniques.

Week 1 - Exploring supervised learning fundamentals

This week, you will develop a strong foundation in supervised learning, progressing from the basics to advanced techniques like polynomial regression and logistic regression. They will also gain practical insights into using activation functions effectively for improved model performance.

Week 2 - Deciphering patterns with neural networks

This week, you will explore neural networks as universal function approximators. There is a specific focus on deep learning and how cascading hidden layers enhance the modelling of complex decision boundaries. You will understand how to predict outcomes through forward propagation and optimise models through backward propagation and gradient calculations. Finally, you will explore the power of frameworks, particularly TensorFlow, to intuitively create neural networks.

Week 3 - Mastering TensorFlow to optimise SL models

This week covers key concepts and practical aspects of utilising TensorFlow for building and tuning supervised learning models. The content covers various layers, model architectures, evaluation metrics, hyper-parameter tuning techniques, and optimisation algorithms.

Week 4 - Optimising model generalisation

This week covers advanced techniques for optimising model generalisation and mitigating overfitting in supervised learning. The focus is on concepts such as bias and variance trade-off, addressing overfitting and underfitting, regularisation techniques including L2 regularisation and dropout, as well as employing callbacks, like early stopping, to enhance model performance.

Week 5 - Mastering decision trees and ensemble techniques

This week covers decision trees and ensemble methods, exploring their power in both classification and regression tasks. There is a focus on fundamental decision tree construction, model training techniques, the learning process, and advanced ensemble methods including bagging, boosting, and the random forest algorithm

Week 6 - Applying advanced modelling techniques

This week focuses on advanced model tuning strategies, AdaBoost, XGBoost, and the interpretation of model decisions through feature importance techniques like SHAP (SHapley Additive exPlanations). Finally, you will complete a 12-hour project activity that encompasses the supervised learning techniques covered throughout the course, enabling you to apply your acquired knowledge to a comprehensive real-world challenge.

Example Project

Classification or regression analysis?

Total duration: 8 hours

What? Explore an organisational scenario, define the problem, and perform relevant analyses.

Why? Embedding selection and problem-solving skills in the context of comparing the methodologies within organisational scenarios will continue to embed the real-life application of key concepts and relevant methodologies.

How? You will be provided with a scenario and will analyse the context to identify the problem. You will then define whether it is a regression or classification problem, select the appropriate methodology, evaluate the model, and summarise the outcome. The output will include an evaluation and report.

This eight-week course covers two of the most critical advanced data science techniques for application in business – natural language processing (NLP) and time series analysis.

Driving many of the recent advancements in AI, NLP offers almost unlimited potential when utilised in the world of business. The ability to break down natural language into data opens up a world of analytical possibilities.

Equally, every business in the world needs to forecast something – whether that’s sales, resources, or other metrics. Time series forecasting offers business solutions to a wide range of challenges, regardless of the business’s domain or the team’s function.

With plenty of time dedicated to hands-on project-based work that will look impressive in any data scientist’s portfolio, you will build upon skills developed earlier in the programme to cover both topics in detail. An in-depth knowledge of these advanced techniques will set you up for senior data science roles in a wide range of sectors.

Course 3 learning outcomes

By the end of this course you will be able to:

• Critically evaluate the application of NLP in a range of organisational contexts
• Leverage expertise in NLP modelling to create solutions to complex organisational challenges
• Master a range of time series forecasting techniques in real-world contexts
• Generate precise forecasts and extract valuable insights from temporal data, contributing to informed decision-making and enhanced operational efficiency in real-world applications.

Week 1 - Preprocessing for natural language processing (NLP)

This week covers essential concepts and techniques in natural language processing (NLP) preprocessing, sentiment analysis, and word embedding vectors. Preprocessing involves preparing text data by removing noise, stemming, lemmatising, tokenising, and analysing n-grams. Sentiment analysis explores using tools like spaCy and NLTK to assess the emotional tone of text. You will then explore embedding vector methods like one-hot representation, word2vec, GloVe, and, finally, how to calculate cosine similarity for representing and comparing words.

Week 2 - Building a basic NLP model

This week, the focus is on applying basic methods like Bag of Words (BoW) and term frequency-inverse document frequency (Tf-idf) to analyse the text corpus. You will then move on to recurrent neural networks (RNNs), including both GRU and LSTM, and discover their practical applications in NLP. Finally, you will advance your skills by building and fine-tuning NLP models

Week 3 - Working with transformer models

This week, you will explore transformer architecture, covering self-attention, multi-head attention, and positional encoding. You will explore and apply advanced models like BERT and GPT for text classification and generation tasks. The focus extends to developing chatbots with NLP capabilities and tackling tasks like topic modelling/text summarisation. The goal is to equip you with essential NLP skills and the ability to address real-world challenges effectively.

Week 4 and 5 - Portfolio focus and consolidation

These two weeks focus on starting two projects that consolidate and apply knowledge gained in the previous three weeks, allowing you to apply NLP for text summarisation and modelling in a real-life context.

Week 6 - Performing time series analysis

This week’s focus is on time series data analysis, essential for understanding and forecasting temporal trends in various domains. The content covers the fundamental characteristics of time series data. You will explore crucial statistical tests for model diagnostics and selection. Additionally, you will explore techniques for handling time series data, including decomposition and transformations, to prepare the data for accurate modelling and forecasting.

Week 7 - Exploring time series forecasting models and methods

This week is dedicated to mastering time series forecasting techniques, with a focus on exponential smoothing, ETS models, the Holt-Winters method, ARIMA models, and advanced modelling techniques. By understanding these methods, you can make precise forecasts, improve decision-making, and optimise resource allocation in various domains.

Week 8 - Applying machine learning and deep learning techniques

This week focuses on utilising ML and DL techniques for time series analysis. You will explore how to leverage the power of both ML and DL statistical techniques by using them in conjunction to form a hybrid model. This will enhance the overall predictive power of the model. You will preprocess, model, and forecast data effectively, arriving at solutions for real-world business challenges and enhancing operational efficiency.

Weeks 9 and 10 - Consolidation and topic project

In the final weeks, you will consolidate and apply your new knowledge of time series analysis with an in-depth 40-hour project.

Example Project

Using time series analysis for sales and demand forecasting

Total duration: 2 weeks (±40 hours)

What? Focus on utilising time series analysis techniques to forecast sales and demand. It involves the analysis of historical sales data to make accurate predictions of future sales trends and demand patterns.

Why? Enhance decision-making by providing reliable sales and demand forecasts to help organisations reduce costs, maximise profits, and optimise resource allocation..

How? You will analyse the employer scenario and explore the data set. You will then clean historical sales data, perform time series analysis using ARIMA, DL, or hybrid methods, and validate the models to ensure their accuracy in predicting future sales and demand patterns. The output will include an evaluation and report.

You will explore the future of data science before demonstrating your skills developed throughout the programme with a live business project run by a real employer.

To begin, you will consider the business and ethical implications of Generative AI, ChatGPT and other large language models (LLMs) in the context of the Data Science profession.

You will finish the programme with a project to showcase the range of skills and competencies gained throughout the course. You will collaborate effectively within a cross-functional team, using multidisciplinary approaches to solve complex real-world problems set by leading industry partners and create strategies that maximise potential business value.

Google Colab, Jupyter Notebook and GitHub

Programming Language: Python

Personalised coaching

Benefit from six 1:1 coaching sessions that provide tailored career advice, aiming to enhance your professional growth in data science. The sessions focus on your unique career aspirations, tackling your questions and equipping you with practical strategies for success.

Group sessions

Participate in four group sessions that address essential career topics, such as how to set meaningful career goals and stay on track, managing your mindset for career success, navigating imposter syndrome with confidence, and team effectiveness. These sessions offer a shared learning experience, enabling you to gain insights from peers while understanding the broader context of their career journeys.