Txt2SQL : SQL-queries from Natural Language Questions and its Practical Business Applications

Sammanfattning: The area of research regarding how to translate Natural Language Questions into SQL-queries has evolved in the last years with developments in Machine Learning algorithms as well as with the emergence of new datasets. This report aims to give insights into the complexities of the task and attempts to evaluate the possible areas of use in a practical context of modern-day corporations. The implementation suggested in this report gave inferior results to the state-of-the-art but illuminated several aspects that, together with a qualitative study, showed that the implementation and usefulness in a practical context are not obvious and require further research.