Consulting Service
Personalized support, from methodological design to the performance of exploitable results.
Methodology
Assistance in study design, choice of sampling plans, definition of models, definition of variables and measurements.
- Documentation and transparency of choices
- Observational and experimental studies
- Randomization, controls, analysis plans
Data preparation
Cleaning, structuring, harmonizing and integrating your data before analysis.
- Preparation for causal and Bayesian analyses
- Creating data pipelines
- Documentation and reproducible scripts
Support for articles and theses
Revision of the Statistics section, method recommendations, assistance in drafting results and interpreting.
- Methodological journals for scientific journals
- Doctoral and Master theses
Causal data analysis
Modeling of treatment effects, causality under observation conditions, reduction of selection bias.
- Predictor adjustment models (regression, matching)
- Propensity score models, difference-of-differences
- Prudent interpretation of causalities
Bayesian modeling
Individual courses or sessions or in groups adapted to your needs on statistics, causality or Bayesian models.
- Hierarchical and multilevel models, like Mixed Marketing Model
- Visualization and communication of post-post distributions
Training and coaching
Individual courses or sessions or in groups adapted to your needs on statistics, causality or Bayesian models.
- Practical workshops with Python/R
- Preparing to analyze your own projects
- Support for adoption of new methods
Why Bayesian statistics and causal methods?
Bayesian statistics are more intuitive and have lessa priori that classic frequentist statistics, although theIntuition wants us to believe otherwise. In addition, they allowobtain answers generatively as much whenThere are differences only whenhe nthere is none. The classic statistics are content to give an answer in case of differences. Otherwise, the decision is suspended. This greatly limits discoveries and leads to well-known dubious scientific behaviors now.
With regard to causal methods, firstly, they push to conceptualize experiments and analyzes more finely because of the representation of thean effect. Then, these methods allowPerform virtual experiments where it doesis not possible in the real world, whether for ethical or statistic reasons (random impossible sampling). Finally, they provide indications on what sis produced, what could have happened (what if…?) and on what will happen.
Indicative Rates
Rates are indicative and may vary depending on complexity, duration and volume of work. Personalized quotes are offered on request.
Students
from 70 CHF/h
For methodological help, courses, interpretation, analytical proofreading and occasional support.
- Choice of statistical test and analysis plan
- Assistance in interpreting the results
- Short formats or targeted packages
Individuals
from 130 CHF/h
For research projects for individuals or academics. Support and analysis of non-commercial data
- Statistical analysis, study design and reporting
- Support SAS, R, Python, SQL as needed
Company
from 150 CHF/h
For freelancers, firms or SMEs requiring advanced modeling or with a high business challenge.
- Causal inference, Bayesian models, Expert audit
- Strategic support
- Implementation of AI systems (Machine Learning and LLM)
No more Status Quo with your data
Contact
Useful information
Working area:
Switzerland (possible remote support)
Languages:
French, English, Italian (basic knowledge) and German (basic knowledge)