Big Data and AI for Banking
Banking data at scale, from storage through analysis to models and dashboards
About this course
A five-module course on working with banking data at scale and putting AI on top of it. It starts with why banking generates the data volumes it does and what distributed storage and NoSQL databases are actually for, then moves through customer-service chatbots and dataset management, data cleaning and exploratory analysis, machine learning on large datasets, and finally visualisation and a full case study.
Every exercise uses synthetic banking datasets generated for teaching — transactions, accounts, customer records, support tickets. No real customer data is used, referenced or required at any point.
Five modules, each with a practice brief and a self-check, ending in an end-to-end banking case study that joins big data, AI and visualisation.
What you’ll be able to do
- Explain what makes banking data a big-data problem, and name the use cases where scale genuinely changes the approach
- Set up and query a NoSQL database, and say when a document or wide-column store beats a relational one for a banking workload
- Build a simple intent-classification chatbot for customer service, and manage its dataset with real versioning
- Clean a messy banking dataset defensibly, recording every decision that changed a number
- Run exploratory analysis and produce the findings that would actually change a decision
- Train and evaluate a machine learning model on a banking problem, with the class imbalance and cost asymmetry handled honestly
- Build visualisations and dashboards that answer a stated question rather than display everything available
- Carry one case study from raw data through to a presented result
Before you start
What this course assumes you can already do.
- Working Python: pandas, reading and writing files, writing a function.
- Basic SQL — joins and aggregation.
- Enough banking domain familiarity to recognise what a transaction, an account and a customer record contain.
- A machine you can install and run database software on, or a cloud environment you control.
- No distributed-systems background required.
The curriculum
Open any module to see its objective, every teaching topic, the practice brief and the checkpoint.
Recaps and reference material are included in the LMS.
PART 1Foundations and storage
01Big Data in Banking and Foundational Concepts
Understand what actually makes banking data a big-data problem — volume, velocity, regulatory retention and the joins nobody can avoid — and get hands on with distributed storage and a NoSQL database so the choice is informed rather than fashionable.
- Understanding Big Data for BankingConcept
- Use cases of Big Data in Banking (fraud detection, risk assessment, etc.)Concept
- Introduction to Hadoop and configuring it for banking useHands-on
- NoSQL databases with Big Data (MongoDB, Cassandra)Concept
- Practical session: Setting up a simple NoSQL databaseHands-on
Justify the storage choice for one banking workload
Take one concrete banking workload — a fraud-screening feed, a regulatory retention archive, a customer 360 view — and decide where its data should live. Then set up a NoSQL database and load synthetic data into it to test your reasoning.
Knowledge checkpoint with explained answers
PART 2Conversational AI and datasets
02Building Chatbots and Dataset Management
Build a working customer-service chatbot from your own intent dataset, and treat that dataset as the engineered artefact it is — versioned, reviewed and traceable when the behaviour changes.
- Introduction to chatbot development for customer service in bankingConcept
- Building datasets for chatbots (text, intent classification, etc.)Hands-on
- Publishing datasets to production and managing dataset versionsConcept
- Practical session: Building and deploying a basic chatbot using NLP librariesHands-on
Build the intent dataset, then version it like code
Build an intent-classification chatbot for a narrow banking scenario — card blocking, balance queries, dispute raising. The assessed artefact is as much the dataset as the bot.
Knowledge checkpoint with explained answers
PART 3Analysis and modelling
03Data Analysis, Mining and Cleaning
Clean a genuinely messy banking dataset and defend every decision that changed a number — then run analysis that produces findings someone would act on, not a gallery of charts.
- Steps in data analysis and the importance of data cleaningConcept
- Data mining techniques and tools for Big DataConcept
- Exploratory Data Analysis (EDA) with Python and common data visualization toolsHands-on
- Practical session: Cleaning and analyzing a banking datasetHands-on
Clean a banking dataset and defend every decision
Work a deliberately messy synthetic banking dataset: duplicate transactions, impossible dates, inconsistent merchant names, missing balances, currency mixed into a numeric field. Clean it and keep a full decision log.
Knowledge checkpoint with explained answers
04AI and Machine Learning for Big Data
Train a model on a banking problem where the classes are badly imbalanced and the two error types cost very different amounts — and make the architecture and model choice for the reason you can state.
- Introduction to AI/ML in Big Data applicationsConcept
- Using Big Data with AI: architecture and model selectionConcept
- Advanced Python for AI & ML (TensorFlow, PyTorch)Hands-on
- Practical session: Building a basic machine learning model using Big DataHands-on
Model an imbalanced banking problem honestly
Train a model on a synthetic banking problem with severe class imbalance — fraud or default, where under two percent of rows are positive. Then evaluate it the way a risk function would.
Knowledge checkpoint with explained answers
PART 4Visualisation and case study
05Case Studies, Visualisation and Practical Use Cases
Turn the whole course into one delivered thing: a use case worth doing, a dashboard that answers a stated question, and an end-to-end case study you can present and defend.
- Creating effective use cases in banking with Big Data and AIConcept
- Data visualization techniques and tools (Power BI, Tableau)Hands-on
- Hands-on: Building an end-to-end case study for Big Data, AI, and visualization in bankingHands-on
Carry one case study all the way through and present it
Choose one banking use case and take it from raw synthetic data to a presented result: cleaned data, a model or analysis, a dashboard, and a case for doing it.
Knowledge checkpoint with explained answers
One banking use case, raw data to presented decision
An end-to-end case study joining big data, AI and visualisation on a banking problem. Built entirely on synthetic banking data.
Use case and data platform — Decide what is worth doing, and where the data will live.
Cleaning and exploration — Make the data defensible before you model anything on it.
Model or analysis — Build the thing, and evaluate it as a risk function would.
Conversational surface (where it fits) — If the use case has a customer-service surface, build it properly.
Visualisation and presentation — Answer one question well, and say what you do not trust.
Next course
Tell us about yourself or your team and we’ll let you know how and when you can take this course.