Statistical & Analytical Consulting
From data to clarity and decisions.
A research-focused consulting group based in Israel, working internationally. We help transform complex datasets and statistical models into clear analytical insights.
We work with clients in Israel and worldwide, supporting both one-time analytical tasks and long-term research collaborations.
About
Leonov Analytics is led by Nickolay Leonov, Ph.D., an applied mathematician with more than fifty years of experience in probability theory, optimization, applied statistics, and interdisciplinary research. He has worked with medical, sociological, marketing, and business data, combining mathematical thinking with practical constraints.
Depending on the project, we involve a global network of external experts in applied statistics, sociology, and marketing research. The team configuration is adapted to each task.
Services
Our work spans several areas. They can be combined into an analytical framework tailored to your needs.
Research consulting (medicine, sociology, marketing)
Methodological and analytical support for empirical research where both domain expertise and a clear analytical methodology are required.
- Questionnaire and study design, operationalization of concepts
- Choosing appropriate metrics and endpoints
- Analysis of observational and experimental data
- Support for publications, theses, and reports
Mathematical modelling & simulation
Building models for complex systems where experiments are costly or infeasible: simulation, exploration of scenarios, and robustness checks.
Python tools & LLM-based automation
Practical tools for data analysis and automation: scripts, analysis templates, and LLM-based utilities for classification, automatic reports, or model interpretation.
Advanced methods & research
In addition to widely used statistical tools, we apply analytical methods developed within our own structural–approximation framework. This approach focuses on achieving clear interpretability and aligning analytical structure with the substantive structure of the data.
The framework assumes that both the input information and the research goals should be organized on a shared methodological foundation while allowing for a wider variety of mathematical structures than those used in classical statistics.
One example is the Parliament method: a family of techniques for constructing relatively small sets of representative objects that summarize the population under study. Depending on the research goal, different variants of the method highlight different types of representativeness—ranging from generalizations of familiar clustering methods such as k-means to conceptually distinct approaches aimed at stability, conservativeness, or compromise.
These methods are used when questions of interpretability, robustness, or structural clarity are central to the analysis, or when standard tools provide unstable or insufficiently meaningful solutions.
Case studies
Below are illustrative examples of typical projects. More detailed descriptions can be provided on request, subject to confidentiality constraints.
Oncology: survival analysis
Goal: evaluate prognostic factors and compare treatment strategies based on time-to-event data.
Approach: careful data cleaning, selection of appropriate survival models, treatment of censoring, and sensitivity checks.
Outcome: a clear statistical report with interpretable conclusions and visual tools for the clinical team.
Sociological research: segmentation
Goal: divide respondents into groups based on attitudes and behaviour to inform communication strategies.
Approach: variable selection, clustering methods, stability checks, and joint interpretation of types with stakeholders.
Outcome: a clear typology, profiles of each group, and recommendations for further research and practical work.
Business & operations: forecasting
Goal: build demand or workload forecasts under uncertainty to support resource planning.
Approach: analysis of historical data, selection and calibration of appropriate models, scenario analysis, and forecast quality assessment.
Outcome: a practical forecasting tool and recommendations on how to integrate it into decision-making processes.
Contact
For an initial contact, a short description of your data and question is usually sufficient. We can then propose several options for collaboration.
Location
Based in Israel, working online with clients in Israel and worldwide.
A possible structure of the first message
You may briefly mention:
- Who you are and what your field is.
- What data or research question you have.
- What kind of result you would like to obtain.
- Any constraints (timeline, format, budget considerations, etc.).