AKALYSIS

Longitudinal data analysis

Longitudinal data analysis for repeated, evolving and time-based data

For projects where the structure of time matters and simple cross-sectional thinking is not enough.

Longitudinal data creates opportunities for richer analysis, but it also raises harder questions: correlation within subjects, changing exposures, attrition, irregular follow-up, competing risks, and what exactly should be interpreted over time. AKALYSIS helps teams work through those problems with methods that match the actual data-generating structure.

Led by Dr. Andrew Kingston, statistician, epidemiologist, data scientist and educator. BSc (Hons) MSc PhD CStat SFHEA.

Area of Focus

Repeated-measures structure

Choosing between mixed models, GEE-style thinking, transition approaches, survival frameworks, and other time-aware strategies.

Area of Focus

Time-varying complexity

Handling changing covariates, follow-up patterns, dropout, and the consequences of irregular observation schedules.

Area of Focus

Interpretation over time

Making sure the model answers the longitudinal question you actually care about, not just a convenient approximation.

When this is a good fit

You have repeated measures, multiple waves, follow-up visits, or panel data.

Attrition, missingness, or timing differences are likely to distort the analysis.

You need help deciding whether to model change, transitions, trajectories, or events.

The question involves ageing, progression, service use, outcomes over time, or changing exposures.

Typical analytical work

  • Linear and generalised mixed models
  • Repeated-measures strategy and design review
  • Attrition, dropout, and missing-data considerations
  • Time-varying covariate handling
  • Longitudinal interpretation for papers, reports, and plans

Why AKALYSIS

Statistical rigour with epidemiological and data science depth

AKALYSIS is designed for projects where the analysis needs to be methodologically coherent as well as useful. That means careful thinking about design, model choice, interpretation, bias, uncertainty, and how results will be challenged by collaborators, reviewers, decision-makers, or the real world.

Book a Free Consultation

Book a free initial call to discuss your data and what rigorous analysis would actually look like for your project. No obligation, no sales pitch.

Free initial consultation, no obligation.