Statistical analysis
Turn your data into actionable findings.
Our statistical analysis services help you draw the greatest possible value from your data. Administrative data, survey results or complex databases: our team of experienced statisticians applies advanced methods to bring out trends, relationships and hidden patterns — and to turn figures into clear accounts and strategic recommendations.
Why choose our analyses?
Advanced technical expertise
Regression, statistical modelling, multivariate analyses, time series, machine learning.
Visualisations that carry
Charts and interactive dashboards that make your data accessible and convincing.
Interpretation set in context
Beyond the figures, expert interpretation that takes your organisational context into account.
Scientific rigour
High academic standards, full documentation of methods and robustness testing.
Our statistical analysis services
Descriptive analyses
Clear summary and presentation of your data: central tendency, dispersion and distributions.
- Descriptive statistics and frequency tables
- Analyses of trends over time
- Population profiles and segmentation
- Cross-tabulations and bivariate analyses
Inferential analyses
Hypothesis testing and parameter estimation, so your results can be generalised.
- Comparison of means (t-test, ANOVA)
- Tests of association (chi-square, correlations)
- Confidence intervals and margins of error
- Non-parametric tests
Statistical modelling
Predictive and explanatory models to make sense of complex relationships between variables.
- Linear and logistic regression
- Multilevel and hierarchical models
- Survival analysis and duration models
- Time series (ARIMA and others)
Multivariate analyses
Advanced techniques to explore latent structures and reduce complexity.
- Principal component analysis (PCA)
- Exploratory and confirmatory factor analysis
- Cluster analysis
- Discriminant analysis
Our analytical approach
Exploration and cleaning
Assessment of data quality, treatment of missing values, detection of outliers.
Exploratory analyses
Descriptive statistics, preliminary visualisations and identification of patterns.
Modelling and testing
Suitable methods, validation of assumptions, robustness testing and interpretation.
Communicating the findings
Reports, visualisations and presentations suited to your audience, with concepts explained in plain language.
Examples of illustrative engagements
These examples describe illustrative engagements: they show how we work. They report no actual engagement and name no client. Engagements we have actually carried out remain our clients’ property and are not disclosed.
Identifying the factors associated with reoffending
An organisation holds several years of case files and wants to know which factors are genuinely associated with risk, so that it can direct its resources better.
Modelling on historical data (logistic regression or survival models, depending on the question), with cross-validation and sensitivity analysis. A model of this kind generally calls for several thousand case files.
- Associated factors, with their significance and effect size
- Limitations of the model stated explicitly
- Documented, reproducible risk profiles
Comparing the effectiveness of several programmes
A funder wants to know which of its programmes produces the most effect for each dollar invested.
Quasi-experimental analysis with propensity score matching, followed by a comparison of effects. The method calls for comparison groups that are large enough and for indicators shared across programmes.
- Effects compared, presented with their confidence intervals
- Cost-effectiveness analysis for each programme
- Profiles of the people who benefit most from each programme
Anticipating peaks in demand
A service wants to plan its staffing from the history of its calls rather than from intuition.
Time series analysis: seasonal decomposition, detection of breakpoints, and a short-term forecasting model together with its margins.
- Weekly and annual seasonality documented
- Forecasts accompanied by their uncertainty
- Monitoring dashboard handed over to the client
Tools and software
- R and RStudio (advanced analyses and visualisations)
- SPSS Statistics (standard analyses and reporting)
- Python (machine learning and big data)
- Stata (econometric and panel data analyses)
- SAS (complex analyses and large databases)
- Tableau and Power BI (interactive dashboards)
Deliverables included
- Analysis report with methodology and interpretation
- Publication-quality tables and charts
- Cleaned data and analysis scripts (on request)
- Interactive dashboards or custom visualisations
- Presentation of the key findings
- Training session on interpreting the results
Put your data to work
You hold data but are not sure how to draw value from it? Find out what your data can reveal.