AQT

University of Texas at Dallas | United States

Hiranya Garbha Kumar | Shalin Anilkumar Amin

Databases are indispensable for almost all companies these days. They provide innumerable features and advantages to a company. As such, accessing or even being able to use the database is a critical skill requirement for people working with it. AQT aims at removing this skill requirement once and for all by translating simple English questions into database queries and also (possibly) performing query optimization to achieve results on par or even better than what a normal employee can. This brings down the skill requirement to access and use a database much lower, allowing even employees without technical skills to directly use the database without relying on a member of the technical team! This not only speeds up different processes but also increases efficiency of the entire team. And the bot does all of this on a server-less system using lambda functions (explained later) which give results much faster than traditional databases. ------------------------------------------------------------------------------------------------------------------------------------------ An example of a query using AQT:- how many residential properties do we have with roof permits issued after 1/1//2002 in the state of Florida on homes built before 2002 ? SQL conversion: select count(distinct residential_property) from xyz.csv where true and preferdate >= date('2002-1-1') and yearbuilt <= 2002 and roof=1 and state = 'fl' The SQL query will get executed on company database using lambda functions and the generated result will be returned to our chatbot UI within few seconds.