Inside the system
Bring your dataset and define what you want to understand. OptiMod checks the data, evaluates suitable methods and carries out the analysis—with findings, uncertainty and a Scientific Report you can inspect.
No coding. No statistics expertise required.
01 / The analytical engine
The question defines the analytical objective. Variable types, data quality and relationships narrow the field of candidate methods. Mathematical requirements and diagnostic evidence inform the final choice.
Check variable types, completeness and dependence. Identify information that would reveal the outcome in advance.
Establish which methods are eligible, fit candidates and compare their assumptions and diagnostics against the objective.
Connect the estimates to the chosen method, its uncertainty and the limitations that matter for interpretation.
OptiMod’s core algorithms were developed in-house, building on several years of research at Leibniz University Hannover. Scientific knowledge becomes part of the workflow, from the first data check to the final report.
Explore the scientific foundation02 / Statistical expertise, built in
A simpler interface is only part of the advance. OptiMod brings scientific decision logic into the analysis itself, so people can investigate their data without having to become statistics specialists.
A conventional statistics workflow
Statistical software provides the methods. The person using it still needs to determine what fits the data, which assumptions to examine and how to interpret the findings.
With OptiMod
You contribute the question and knowledge of your field. OptiMod carries out the statistical workflow and keeps the reasoning connected to the results.
04 / The hybrid architecture
OptiMod combines scientific knowledge, mathematical calculations and targeted AI assistance. Each contributes to an analysis that is efficient to run and clear to understand.
The analytical foundation
Scientific requirements and numerical evidence guide the analytical path, from the methods that fit the data to the estimates they produce.
Support where it adds value
AI supports data understanding and efficient analysis, helping turn complex findings into explanations that stay connected to the calculations.
For scientific and business-critical decisions, the bottleneck is no longer producing an answer, but trusting it.
05 / Your analysis, delivered
Export your Scientific Report as a PDF. It documents data preparation, method selection, model assumptions, the named tests used to check them and the results. Review, share and retain the analysis independently of your access to OptiMod.
Example extract · simulated study
Change in symptom score · lower is better
The example package contains the synthetic dataset, calculation script and full numerical results.
Download the example package (ZIP)In this independently computed example using public Capital Bikeshare data, both count models were fitted to the first 547 days. Predictions were evaluated on the next 184 days using recorded weather.
| Candidate | Training AIC | Holdout MAErides/day |
|---|---|---|
| Poisson | 66,197.6 | 689 |
| Negative binomial (NB2) | 8,815.7 | 863 |
Lower values are better for both measures. AIC assesses model fit relative to complexity. Mean absolute error (MAE) measures prediction error.
Association analysis. The Poisson fit showed substantial extra variation. NB2 had the lower training AIC and was selected to study associations in this example.
Prediction error. On the 184 held-out days, Poisson predictions were off by 689 rides per day on average, compared with 863 for NB2. This comparison used observed weather.
06 / Beyond the workspace
Use OptiMod directly or explore bringing its analysis capability into your own software, data platform or AI assistant. The interface can change; the need for scientifically grounded, traceable findings stays the same.