Utilizing artificial intelligence for data management and data visualization

1 . A method, comprising:

receiving, via a text input box associated with a user interface of a client device, an initial natural language prompt associated with a user;

generating a first dashboard based on the initial natural language prompt utilizing a machine learning model and user settings;

receiving, via the user interface of the client device, one or more follow-up prompts, wherein the follow-up prompts refine the initial query, adjust parameters, or apply filters on the fly;

performing a vector search over edges associated with a plurality of knowledge graphs to identify relevant entities to the follow-up prompts, wherein the edges comprise of subject, predicate, and object properties which are indexed through a data warehouse;

based on results of the vector search, determining one or more data catalogs corresponding to one or more data files needed to answer the follow-up prompts;

providing the one or more determined catalogs to a large language model;

utilizing the large language model to generate a structured query language (SQL) command to obtain the one or more data files needed to answer the one or more follow-up prompts;

executing the command to obtain the one or more data files;

applying one or more security policies to the one or more obtained data files using column-level access permissions stored in the one or more data catalogs to constrain results of executing the command to only data columns the user is authorized to access, and masking or anonymizing data to prevent the leakage of private user information; and

generating, at a display of the client device, one or more subsequent dashboards utilizing the machine learning model and the user settings associated with the user based on results of executing the command.

2 . The method of claim 1 , wherein the machine learning model is a reinforcement learning model, a supervised learning model, an unsupervised learning model, or a semi-supervised learning model.

3 . The method of claim 1 , wherein the user settings are updated each time a dashboard is generated for the user.

4 . The method of claim 1 , wherein the user settings indicate types of queries the user is asking, a type of information the user is interested in viewing in a dashboard with respect to different query types, types of widgets used to visualize the information, a configuration of the widgets on the dashboard, and/or size of the widgets on the dashboard.

5 . The method of claim 1 , further comprising receiving feedback on the first dashboard.

6 . The method of claim 5 , further comprising updating the user settings associated with the user based on the received feedback.

7 . The method of claim 1 , wherein the one or more follow-up prompts include natural language.

8 . The method of claim 1 , further comprising utilizing the one or more data catalogs corresponding to the one or more data files to interpret the data associated with the one or more data files.

9 . The method of claim 8 , wherein utilizing the one or more data catalogs corresponding to the one or more data files to interpret the data associated with the one or more data files includes normalizing the data associated with the one or more data files.

10 . The method of claim 8 , wherein generating the one or more subsequent dashboards includes visualizing the obtained data associated with the one or more follow-up prompts based on the one or more data catalogs corresponding to the one or more data files.

11 . A system, comprising:

a processor configured to:

receive, via a text input box associated with a user interface of a client device, an initial natural language prompt;

generate a first dashboard based on the initial natural language prompt utilizing a machine learning model and user settings;

receive, via the user interface of the client device, one or more follow-up prompts, wherein the follow-up prompts refine the initial query, adjust parameters, or apply filters on the fly;

perform a vector search over edges associated with a plurality of knowledge graphs to identify relevant entities to the follow-up prompts, wherein the edges comprise of subject, predicate, and object properties which are indexed through a data warehouse;

based on results of the vector search, determine one or more data catalogs corresponding to one or more data files needed to answer the follow-up prompts;

provide the one or more determined catalogs to a large language model;

utilize the large language model to generate a structured query language (SQL) command to obtain the one or more data files needed to answer the one or more follow-up prompts;

execute the command to obtain the one or more data files;

apply one or more security policies to the one or more obtained data files using column-level access permissions stored in the one or more data catalogs to constrain results of executing the command to only data columns the user is authorized to access, and mask or anonymize data to prevent the leakage of private user information; and

generate, at a display of the client device, one or more subsequent dashboards utilizing the machine learning model and the user settings associated with the user based on results of executing the command; and

a memory coupled to the processor and configured to provide the processor with instructions.

12 . The system of claim 11 , wherein the one or more follow-up prompts include natural language.

13 . The system of claim 11 , wherein the processor is configured to utilize the one or more data catalogs corresponding to the one or more data files to interpret the data associated with the one or more data files.

14 . The system of claim 11 , wherein generating the one or more subsequent dashboards includes visualizing the obtained data associated with the one or more follow-up prompts based on the one or more data catalogs corresponding to the one or more data files.

15 . A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:

receiving, via a text input box associated with a user interface of a client device, an initial natural language prompt associated with a user;

generating a first dashboard based on the initial natural language prompt utilizing a machine learning model and user settings;

receiving, via the user interface of the client device, one or more follow-up prompts, wherein the follow-up prompts refine the initial query, adjust parameters, or apply filters on the fly;

performing a vector search over edges associated with a plurality of knowledge graphs to identify relevant entities to the follow-up prompts, wherein the edges comprise of subject, predicate, and object properties which are indexed through a data warehouse;

based on results of the vector search, determining one or more data catalogs corresponding to one or more data files needed to answer the follow-up prompts;

providing the one or more determined catalogs to a large language model;

utilizing the large language model to generate a structured query language (SQL) command to obtain the one or more data files needed to answer the one or more follow-up prompts;

executing the command to obtain the one or more data files;

applying one or more security policies to the one or more obtained data files using column-level access permissions stored in the one or more data catalogs to constrain results of executing the command to only data columns the user is authorized to access, and masking or anonymizing data to prevent the leakage of private user information; and

generating, at a display of the client device, one or more subsequent dashboards utilizing the machine learning model and the user settings associated with the user based on results of executing the command.