Skip to content
ezyStat

Research, Statistics & Data Intelligence

Turning data into evidence.Evidence into better decisions.

ezyStat provides expert statistical consultancy, data analytics, polling, market research and feasibility studies to help researchers, businesses and organisations make informed decisions.

Explore our servicesDiscuss your project
Illustrative linear regressionScatter plot of 48 simulated observations with a fitted regression line, a 95% confidence band and a wider 95% prediction band, plus a density curve of the outcome on the right.0204060800246810ŷ = 14.5 + 5.15xR² = 0.86, n = 48
Figure 1. Simulated data with a fitted least-squares line, its 95% confidence band (darker) and 95% prediction band (lighter). The curve on the right is a kernel density estimate of the outcome. Illustrative only, not client data.

Six ways we work with data

Each service can stand alone or combine with others. A survey can feed a market study; a market study can feed a feasibility assessment.

Why choose ezyStat?

Rigour first, then clarity. You get the methodological standards of academic research, explained in language that supports a decision.

The value of an analysis lies in the decision it supports, and in being honest about how certain that decision can be.

Our working principle
Advanced statistical expertise
Work is led by a PhD statistician, not a generalist analyst, so methods are chosen for the problem rather than habit.
Rigorous research methodology
Design comes before data. Sampling, measurement and analysis are planned together, and assumptions are written down.
Independent analytical advice
We report what the evidence supports, including when it does not support the hoped-for answer.
Clear, actionable reporting
Results are explained in plain language, with the uncertainty stated and the practical implications drawn out.
Flexible, project-based services
From a single methodology review to a full survey programme, scope and cost are agreed before work begins.
UK and international orientation
Studies in the UK and abroad, with particular interest in Middle Eastern research and multi-country work.
Academic and commercial understanding
Comfortable with peer review and with board papers, and clear about the different standards each requires.

Our research process

The same six stages apply to a small analysis and a multi-country survey. What changes is the depth at each stage.

  1. 1

    Understand the research question

    We start with the decision or question behind the project, not the data.

  2. 2

    Define the methodology

    We agree what will be measured, how, and what would count as a clear answer.

  3. 3

    Design the study

    Sampling, instruments, models and power are specified before any data are collected.

  4. 4

    Collect and validate data

    Data are gathered or assembled, then checked for quality, completeness and bias.

  5. 5

    Analyse and interpret

    Analysis follows the plan, with diagnostics and sensitivity checks on the key results.

  6. 6

    Deliver insights

    You receive a clear report with recommendations, uncertainty and next steps.

Polling and market research, with the uncertainty stated

Quantitative surveys, public opinion research and international market intelligence, designed from the sampling frame up. We manage methodology, sampling, quality assurance and analysis, and coordinate CAWI, CATI and CAPI fieldwork directly or through qualified local partners.

Margin of error at 95% confidence, simple random sample

n = 500
±4.4
n = 1,000
±3.1
n = 2,000
±2.2

Percentage points, for a 50% result.

Polling and survey researchMarket research

“Do you support or oppose the proposal?”

Illustrative poll estimates with 95% confidence intervalsStrongly support: 19%, plus or minus 2.4 points; Tend to support: 32%, plus or minus 2.9 points; Neither: 16%, plus or minus 2.3 points; Tend to oppose: 20%, plus or minus 2.5 points; Strongly oppose: 13%, plus or minus 2.1 points0%10%20%30%40%Strongly support19% ±2.4Tend to support32% ±2.9Neither16% ±2.3Tend to oppose20% ±2.5Strongly oppose13% ±2.1
Figure 2. Simulated responses to a five-point question from a simple random sample of n = 1,000, with 95% confidence intervals. Real surveys also account for weighting and design effects, which widen these intervals. Illustrative only.

Statistical expertise

Methods are chosen for the question and the data, not for familiarity. Work is delivered in the software your team already uses wherever possible.

Figure 3. Inference. Standard normal density with the upper 5% tail shaded, the basis of a one-sided test at α = 0.05.
Figure 4. Forecasting. Seasonal time series with a ten-step forecast and widening 95% interval.
Figure 5. Group comparison. Box plots comparing three groups: median, interquartile range and 1.5 × IQR whiskers.

Methods we use regularly

  • Regression and GLMs
  • Mixed-effects models
  • Bayesian methods
  • Quantile regression
  • Time series and forecasting
  • Survey sampling and weighting
  • Experimental design
  • Multivariate analysis
  • Machine learning
  • Non-parametric methods

Software

R
Modelling, survey analysis, reproducible reports
Python
Data pipelines, machine learning, automation
SPSS
Academic and survey analysis
Excel
Models, scenarios and accessible outputs
Power BI
Interactive dashboards, where appropriate
Quarto / R Markdown
Reproducible technical reports

Led by a mathematical statistician

Arash is a mathematical statistician with more than fifteen years in higher education and research. He has taught statistics, mathematics, research methods and programming at universities in the United Kingdom, New Zealand and Iran, and has supported learners and researchers at many different academic levels.

Dr Arash Ardalan founded ezyStat to give researchers and organisations direct access to that expertise, with the same care for method and explanation he brings to teaching.

  • PhD in Mathematical Statistics
  • MSc in Statistics
  • BSc in Statistics
About ezyStat and its founder

Resources and insights

Practical guides on study design, analysis and reporting are in preparation. The sample size calculator is available now.

  • Survey designPlanned article

    How to determine sample size for a survey

    Margin of error, confidence level, subgroups and the effect of weighting, worked through with examples.

  • Statistical guidesPlanned article

    Choosing the right statistical test

    A practical route from your research question and data type to an appropriate test or model.

  • Research methodologyPlanned article

    Market research versus feasibility studies

    What each one answers, when you need both, and how they fit together.

Try the sample size calculatorVisit the resource centre

Have a research question or data challenge?

Tell us about your project and discover how ezyStat can help you design, analyse and interpret your data.

Request a consultation