MOSS is a clinical data analysis and visualisation tool which makes use of Natural Language Processing and Deep Learning algorithms to automatically translate a natural language (English) question from a clinician to a structured database query. MOSS will then render the output in an aesthetically appealing form (charts, graphs, tables, tokens) and will also allow drill down analysis on these reports. MOSS allows clinicians and trial monitors to reduce their dependence on the data team to do ad-hoc visualisations and analysis of key risk indicators. MOSS will also provide statistical analysis plugins in future releases so that the clinicians will be able to perform basic data analysis.
Clinicians can ask their questions in natural language. MOSS NL2SQL translator will translate the questions into structured database queries. MOSS will then run those queries on the database and visualise the data
NL2SQL models can be trained for any database with a few thousand questions. This training module makes MOSS database agnostic and hence can work across multiple EDC schemas and standard datasets like SDTM and ADAM
MOSS has an aesthetic visualisation model which will render the data based on the choice of the users. In addition to the standard representations like bar charts and pie charts, MOSS also supports complex visualisations like heatmaps and forest plots.
MOSS provides drill down capabilities which will help the users to click on the charts and further drill down the data. This allows the clinicians to go in depth into the data and perform further analysis on the go.