Data science, machine learning (ML) & artificial intelligence (AI)

“So what’s getting ubiquitous and cheap? Data. And what is complementary to data? Analysis.” Hal Varian

We use machine learning across our practice areas to extract insights from data.

Finding patterns in data

Consumption profiles, trade flows, supply chains, networks of institutions, products and individuals: Machine Learning provides us with the tools to find, map and analyse the complex structures contained in data, unlocking new and powerful insights for businesses and policymakers.

Examples:

  • Identifying sub-groups of individuals under financial stress
  • Identifying typical load profiles for industrial and commercial electricity users

Big Data

Smart meters, embedded sensors, earth observation satellites and digital transaction records: we have the tools to store, process and analyse large datasets efficiently and securely.

Examples:

  • Smart meter data on hourly electricity consumption
  • Earth observation (EO) data
  • Large-scale socio-demographic panel data

Predictive modelling

We use powerful machine learning algorithms to select, calibrate and test predictive models, complementing our traditional strengths in econometrics.

Examples:

  • Predicting the adoption of energy saving methods based on survey data
  • Predicting foreign direct investment (FDI) decisions by life science businesses
  • Predicting Willingness to Pay for geospatial information based on an online experiment

Record linkage

Combining data from different sources is a constant challenge in applied economic research: using supervised and unsupervised learning algorithms, we provide a state-of-the-art, accurate and scalable approach to data matching tasks.

Examples:

  • Quarterly deduplication of a customer data base based on company contact information
  • Merging administrative business database at site level

Data visualisation

Helping our clients to make sense of sophisticated statistical models and complex datasets is a challenge we take very seriously. Our approach is to use clear, intuitive visuals that let the data speak.

Examples:

  • Geospatial clustering of meter locations to identify multi-meter facilities
  • Visualisation of complex analytical outputs
  • Interactive dashboards & monitoring tools

Intelligent nudging

Working with rich behavioural data to design intelligent nudges improving outcomes for business and customers.

Examples:

  • Optimising customer loyalty and retention
  • Identifying indicators of consumer vulnerability
  • Impact of personalised offers and pricing

For queries related to the above areas, contact: Moritz Godel, +44 (0) 20 3701 7708, mgodel@londoneconomics.co.uk