Platforms for science tend to replicate paper-based documentation and folder- based organization on the computer screen. The scientific process, from data acquisition to publication, could be greatly improved by recent advances in network theory, artificial intelligence and data mining. Stepping in this direction, we created a scientific database system where the experimental data is homogenized and deep-indexed. This allows the autonomous exploration of data to identify numeric features and correlations in large material datasets. ScienceDesk’s final goal is to create a friendly environment for data ingestion into research cycles, where correlated data, samples, experiments and articles can be efficiently retrieved.