Forecast Analytics builds and integrates predictive models to enhance business performance and enable data-driven decision making.

Each engagement brings its own set of challenges and opportunities and Forecast is able to draw upon an extensive toolkit to address each problem set.

what we do:

We work with clients across all industries to bring them a competitive advantage from their data by facilitating better and faster decision-making using a data-driven approach.

Our work focuses on advanced analytics and is rooted in technical expertise, client experience and delivery. Forecast Analytics team members all hold advanced degrees in fields ranging from statistics to financial mathematics to computer sciences and have backgrounds in a variety of fields including academia, (retail) finance.

Depending on the nature of the engagement, Forecast team members can act in an advisory capacity or as full team members. The scope of the work Forecast is able to engage in includes predictive modelling, customer segmentation, price optimization and many more.

The Forecast Toolkit


Statistical Machine Learning is at the core of almost all projects Forecast works on and forms the basis to uncover patters and inform business decisions.

Statistical Machine Learning


Deep Learning covers a collection of techniques within Machine Learning and Artificial Intelligence that is very successful in problem sets that involve images and natural language.

Deep Learning


Visualisations are crucial in the successful delivery of any analytics engagement. Some engagements also include an explicit visualisation component such as Tableau or PowerBI.

Data Visualisation


An increasing amount of unstructured data is available and this type of data requires a specific set of tools to handle them. Forecast has experience of building classification models.

Natural Language Processing


Part of general notion of Artificial Intelligence, automating repetitive tasks using simple scripting languages can often yield very attractive and quick returns.

Intelligent Process Automation


Unfortunately, many ML projects never make it into production. Forecast have extensive experience of integrating ML models into our client's workflow using cloud resources or custom APIs.

Machine Learning in Production

our team:

Key Players

Reach out to one of our Directors to empower your business performance.

Neil Macdonald


Greg Norman


Paul Van Loon

Head of Analytics, UK

Abdul Rabbani Shah

Analyst, UK

Senior Manager’s Manual to Machine Learning

Machine learning algorithms receive no explicit programming instructions on how to make predictions but instead detect patterns and learn how to make predictions by processing data. In response to new data, models can be re-estimated. Three types of analytics exist, with the latter being the domain of Machine Learning:


  • Describes what happened
  • Commonly employed across industries
  • Business intelligence,  statistical analysis

Predictive (ML domain)

  • Anticipate what will happen
  • Probabilistic in nature
  • Employed by organisations  with a data-driven culture

Prescriptive (ML domain)

  • Relies on predictive models to provide action
  • Employed by leading  data companies

Supervised Learning

Subdomain of machine learning where training data and known outcomes / labels are used to learn the relationship of given inputs to a given output (for instance, how “square footage”, “interest rates” and “no. of bathrooms” predict property prices). Supervised Learning is employed when you know the input data and the behaviour (labels) you want to predict, would like the algorithm to learn the relationship between input and behaviour and would like the algorithm predict behaviour on new, unseen data.

Supervised Learning techniques are plentiful and include Linear Regression, Logistic Regression, Decision Tree, Naïve Bayes, Support Vector Machine, Neural Networks, Random Forest, Gradient-Boosting trees.

It is impossible to only link specific business case to a specific algorithm; a number of different models can often solve the same business problem. On the other hand, the specifics of a given data set might preclude a model from being applicable. Most broadly, problems that fall under the supervised learning umbrella can be either classification (predicting distinct categories) or regression (predicting a continuous output).

  • Predict whether registered users will be willing or not to  pay a particular price for a product
  • Classify fraudulent or suspicious transactions in a credit  card history
  • Produce forecasts for product demand and associated  inventory levels
  • Predict the propensity to default on mortgage payments
  • Predict customer churn for a new product launch

  • Predict the probability a patient joins a healthcare  programme
  • Predict propensity to click on an online banner ad
  • Model sales drivers such as competition prices, distribution,  advertisement, etc.
  • Predict volume of calls for a call centre and produce  required levels of staffing
  • Predict the price of motorbikes using characteristics

Unsupervised Learning

Subdomain of machine learning where data without an explicit output variable. It may explore customer demographic data to identify patterns). Useful when you have large amounts of data without an idea on how to classify the data and want to algorithm to find these patterns.

Unsupervised techniques include predominantly clustering and dimensional reduction algorithms including k-means clustering, hierarchical clustering, recommender system, Uniform Manifold Approximation and Projection, Stochastic Neighbour Embedding.

Deep Learning

Deep learning covers a specific subset of machine learning that requires less data pre-processing and can produce more accurate than traditional ML techniques across several data types and domains such as image classification, facials recognition and natural language processing. Deep learning builds up neural network by stacking multiple layers that learn increasingly complex features of the data at each layer.

case studies:

Latest Case Study

Telecommunications Demand Forecast Model

Purpose The client required a detailed model to forecast daily enterprise service order volumes and their movement through the development and activation process. The client…

contact us:

Forecast Analytics Europe

Neil Macdonald (Director)


+44 7570 961 716


8 St. James’s Square

St. James’s

London SW1Y 4JU


6-8 Dewar Place Lane

Edinburgh EH3 8EF

Forecast Australia

Greg Norman (Director)


+61 435 863437


94 Jones Bay Wharf

26-32 Pirrama Rd

Pyrmont NSW 2009


401 Collins Street

Melbourne VIC 3000