• Data Analytic Thinking
• Business Problems and Data Science Solutions
• Machine Learning Fundamentals
• Statistical Learning
• Assessing Model Accuracy

• The FDP Charter is an industry-led, practical, challenging assessment of financial data science skills. But we begin with the basics - specifically understanding what is Data Science within the context of finance and investing.
• As well as understanding the key applications of Data Science within finance, we also introduce key Machine Learning (ML) and Statistical Learning concepts - and the basics of model validation.

• Regression & Classification
• Logistic Regression & SVMs
• Regularization & Model Selection
• Time Series Analysis & Forecasting

The foundations of financial data science! Covering:

• Time series and econometrics: Time series characteristics and properties for finance, and how econometric models are tailored for economics applications.
• Linear & Logistic Regression: well-understood linear models, with "guard rails" to prevent a model from going really astray & with high interpretability. Understanding these approaches is a foundation of data processing and data analysis - they are ubiquitous in many financial applications.
• Support Vector Machines: the first specific consideration of Machine Learning (ML). A relatively simple method to classify information or run regressions that may offer you some additional insights vs Linear / Logistic regressions.
• Regularization: how to generalize a model to not only explain the past successfully or in detail, but also reliably use the model for prediction and pattern recognition on a go-forward basis.

• Decision Trees & Supervised Learning
• Ensemble Learning (Bagging, Boosting, Random Forests)

• Decision Trees: a hierarchical method of decision-making, which contrasts with typical decision-making in finance.
• Ensemble Methods: popular, well-known Machine Learning (ML) models that use multitudes of Decision Trees such as random forests (Supervised Learning).

• Nearest Neighbor & Clustering
• Probabilistic Models & Naïve Bayes

• Clustering: the first consideration of Unsupervised Learning, where training data is unavailable. Essentially, what are the relationships that exist in the data?
• Classification: one of the functions of Machine Learning (ML) models, to classify data including multi-label classification.
• Naïve Bayes: a method of classification, a relatively simple application of Bayesian methods. Naïve Bayes is primarily data-driven, minimizing subjective views from the model-builder.

• Neural Networks & Deep Learning
• Gradient Descent & Optimization
• Applications in Finance & Reinforcement Learning

• Moving into more sophisticated algorithms, Neural Network applications / frameworks can be Supervised or Unsupervised. Taking inspiration from the working of human neurons, they can bring some important advantages over modelling e.g. inherent parallel processing.
• Neural Networks have unique characteristics. We review the technical details that allow them to more powerfully help predict and recognise patterns, as well as the drawbacks especially around opaqueness due to their multiple layers - as well as randomization elements.
• Reinforcement Learning is covered as the "third category" of Machine Learning (ML), after Supervised & Unsupervised Learning, that operates on the basis that both humans and models can 'learn' based on a system of rewards and punishments. This is helpful in finance as it mirrors the natural consequences of good & bad investing - making money (reward) or losing it (punishment). Unlike Supervised Learning, you can model financial data with minimal training / historic training data.

• Model Performance Metrics
• Model Interpretability
• False Discoveries & Statistical Pitfalls

• One of the most unappreciated aspects of what quant analysis can do on Wall Street! Even on the qualitative side of investing, there is huge importance to model validation, proper testing including back-testing and cross-validation.
• FDP sets out the ideas that investors should take a scientific approach to investing, or other financial models, which means you should try to break your theory (e.g. an investment strategy) through testing. If you can't break it (falsify or prove wrong), then it has a reasonable chance of working. Without this robustness there is significant risk of false positives on investment strategies which can prove hugely costly.
• This is also important for developers and coders who are building these models. This is the most practical topic of the entire curriculum, because it really gets to the core of what should be driving everything you do with financial data.

• Text Representation & NLP Basics
• NLP Applications in Finance
• Large Language Models (ChatGPT, BERT, etc.)

• Two relatively new areas. The first is Textual Analysis - an introduction to non-numerical, unstructured or semi-structured data, as text data is the most common example as we communicate this way. In finance, the human side of the data is incredibly important and so these techniques help us include this in our models.
• Large Language Models (LLMs) are an offshoot of natural language processing. Understanding how natural language is analysed by various models is important, as is understanding how large language models function. FDP Charter will also allow you to understand how these models differ, which is vital if you want to use them to process or organize data in your models.

• AI Ethics & Bias
• AI in Credit Underwriting
• AI Regulation & Governance

• Ethics permeates everything that we do. You must act in an ethical manner. That includes and respecting privacy when you're dealing with certain kinds of data, and regulations as well. Regulations in financial services are typically applied on the strategy and client side, but being aware of them when models are being developed and data is being used is also important.
• In the future we may get specific regulations regarding the usage of different kinds of data as well, as well as on the use of AI.

• Alternative Data & Investment Management
• AI in Securities Markets
• Multiple Hypothesis Testing
• Text Mining in Finance
• Machine Learning in Finance & Backtesting

• Fintech Applications covers many adjacent areas to financial data - such as blockchain, crypto, and payment systems - that have significant technical, legal, regulatory, and ethical implications to how financial data analysts can operate.
• FDP Charter is a practical, focused designation - the "Trade School of Financial Data" - and we believe that you cannot be an effective financial data analyst with only a theoretical understanding. The real-world applications and implications of your work are crucial to becoming a well-rounded financial data professional.