Cognition Shared Solutions, present on the market since 2007, is a company focused on solving data science and big data related problems.We offer wide range of services including:
- assistance in information retrieval,
- (big) data processing and understanding,
- experiments and analytical protocols design,
- statistical analysis and interpretation,
- prediction models and forecasting.
In our work we use a collection of top edge data inspection tools ranging from the classical frequentist inference to modern machine learning methods. We always pay particular attention to uncertainty measurements and effect size interpretation in order to guarantee the most comprehensive and reliable results.
We have gained recognition and trust by preparing, conducting and analysing of web surveys, constructing quantitative models of bio-molecules production, and solving big data classification problems for many international partners, such as: Cognition Shared Solutions LLC (U.S.A), United R&D sp. z o.o. (Poland), or Applied Research Institute for Prospective Technologies (Lithuania).
We can support your company in
HOW DO WE DO IT?
DATA SCIENCE PROFITS
The conclusions and key data features are always summerized in informative and tailored plots and charts.
Reproducible analytic pipelines
Our solutions are always documented by open source code and explanation in a form of RMarkdown reports.
Interpretation of the results
All analyses end with a delivery of comprehensive and conclusive reports and presentations.
Data audit and analysis design
We provide a complex design of the analyses needed for your business that enable cost-efficient collection of data and the most conclusive results. We design questionnaires and surveys and predict the required number of participants.
The first step is always descriptive statistics and exploratory analysis of data, then we perform statistical inference with uncertainty acknowledged by confidence intervals, and outcome significance evaluated by the effect size interpretation (if possible); modern supervised and unsupervised statistical learning methods are applied, if necessary.
Exploratory analysis and data cleaning
The necessary data cleansing is always performed prior analysis, as tidy data dramatically speed downstream data analysis tasks.