Question library for data analytics interface
A question library aids in intuitive analysis of stored data. The question library comprises: 1) a plurality of text questions, 2) a numerical representation (e.g., a vector) of each text question, and 3) a corresponding query in a query language. A numerical vector is generated for a question posed to a database. If a matching library question (based upon vector similarity) is not found, the user receives the original answer. If a matching library question based upon vector similarity is found, the user receives the answer to that library question (with potential modifications). Embodiments may determine similarity by calculating Pearson's coefficient, Spearman's rho, or Kendall's tau. Embodiments may parse the first query to identify constituent elements (measures, dimensions, filters). These entities are extracted and compared to elements of the second question matched within the library, to allow modification of the library query to align with the initial query.
1. A computer-implemented method comprising:
receiving a first text question;
converting the first text question to a first vector;
referencing a second text question stored in a question library associated with a second vector and having a first hierarchy including a topic and an entity;
referencing a mapping between a data model and the second text question to generate a similarity metric between the first vector and the second vector, the data model having a second hierarchy including the entity;
modifying a second query from another query stored in the question library associated with the second text question and the second vector, by,
parsing the first question according to a first expression comprising a discrete named value and a second expression comprising a temporal expression,
generating an inverted index from the first expression,
referencing the inverted index to change a filter in the another query, and
applying machine learning to the second expression such that the first expression is not overwritten in the second query based upon a prior input to mark the first expression in the question library; and
communicating the second query based on the second question for execution upon a data set stored in a database and organized according to the data model.
2. A method as in claim 1 wherein the second query is stored in the question library associated with the second text question and the second vector.
3. A method as in claim 1 wherein the similarity metric comprises a correlation coefficient.
4. A method as in claim 3 wherein the correlation coefficient comprises a non-parametric rank correlation coefficient.
5. A method as in claim 4 wherein the non-parametric rank correlation coefficient comprises Spearman's rho.
6. A method as in claim 1 wherein:
the database comprises an in-memory database;
the question library is stored in the in-memory database; and
the engine comprises an in-memory database engine of the in-memory database.
7. A non-transitory computer readable storage medium embodying a computer program for performing a method, said method comprising:
receiving a first text question;
converting the first text question to a first vector;
referencing a second text question stored in a question library associated with a second vector and having a first hierarchy including a topic and an entity;
referencing a mapping between a data model and the second text question to generate a similarity metric between the first vector and the second vector;
communicating a second query based upon the second text question, for execution upon a data set stored in a database and organized according to the data model, the data model having a second hierarchy including the entity,
wherein the second query is modified from another query stored in the question library associated with the second text question and the second vector, by,
parsing the first question according to a first expression comprising a discrete named value and a second expression comprising a temporal expression,
generating an inverted index from the first expression,
referencing the inverted index to change a filter in the another query, and
applying machine learning to the second expression such that the first expression is not overwritten in the second query based upon a prior input to mark the first expression in the question library.
8. A non-transitory computer readable storage medium as in claim 7 wherein the similarity metric comprises a correlation coefficient.
9. A non-transitory computer readable storage medium as in claim 8 wherein the correlation coefficient comprises a non-parametric rank correlation coefficient.
10. A non-transitory computer readable storage medium as in claim 9 wherein the non-parametric rank correlation coefficient comprises Spearman's rho.
11. A computer system comprising:
one or more processors;
a software program, executable on said computer system, the software program configured to cause an in-memory database engine of an in-memory source database to:
receive a first text question;
convert the first text question to a first vector;
reference a second text question stored in a question library associated with a second vector and having a first hierarchy including a topic and an entity;
reference a mapping between a data model and the second text question to generate a similarity metric between the first vector and the second vector, the data model having a second hierarchy including the entity;
modify a second query from another query stored in the question library associated with the second text question and the second vector, by,
parsing the first question according to a first expression comprising a discrete named value and a second expression comprising a temporal expression,
generating an inverted index from the first expression,
referencing the inverted index to change a filter in the another query, and
applying machine learning to the second expression such that the first expression is not overwritten in the second query based upon a prior input to mark the first expression in the question library; and
communicate the second query based on the second question for execution upon a data set stored in the in-memory database and organized according to the data model.
12. A computer system as in claim 11 wherein the second query is stored in the question library associated with the second text question and the second vector.
13. A computer system as in claim 12 wherein the in-memory database engine is further configured to modify the second query from another query stored in the question library associated with the second text question and the second vector.
14. A computer system as in claim 11 wherein the in-memory database engine is further configured to generate the similarity metric by calculating a correlation coefficient.
15. A computer system as in claim 11 wherein the data model is also stored in the in-memory database.
16. A computer system as in claim 11 wherein the mapping is also stored in the in-memory database.