Conversation based data analysis recommendations

1. A computer-implemented method comprising:

detecting, in a dataset, an insight, the insight comprising a fact about a portion of data in the dataset, the insight maintained in a text-only structured format comprising a plurality of vocabulary terms, each vocabulary term in the plurality of vocabulary terms comprising a field and a plurality of options for values of the field;

determining, by analyzing a natural language input, a context, the context identifying an intent and an entity referenced by the natural language input, the context maintained in the text-only structured format;

computing, for the insight, a relevance score, the relevance score comprising a measure of similarity between an encoded form of the insight and an encoded form of the context, the encoded form of the insight and the encoded form of the context generated according to the same encoding scheme, the encoded form of the insight comprising an encoding of a vocabulary term used in the insight, the encoded form of the context comprising an encoding of a vocabulary term used in the context, wherein the encoding scheme is a one-hot encoding scheme in which each option in the plurality of options for values of the field is represented by a dimension of a vector; and

presenting a selected insight, the selected insight selected from a set of insights including the insight, the selected insight selected based on the relevance score.

2. The computer-implemented method of claim 1 , further comprising:

determining, by analyzing a second natural language input, an updated context;

detecting, responsive to determining that the updated context has below a threshold similarity with the context, a second insight;

computing, for the second insight, a second relevance score; and

presenting a second selected insight, the second selected insight selected from a second set of insights including the second insight, the second selected insight selected based on the second relevance score.

3. The computer-implemented method of claim 1 , further comprising:

computing, for the insight, an interestingness score;

computing, for the insight, a timeliness score; and

combining, into an insight score, the interestingness score, the relevance score, and the timeliness score.

4. The computer-implemented method of claim 3 , wherein the selected insight is selected based on the insight score.

5. The computer-implemented method of claim 3 , wherein the insight score comprises a weighted average of the interestingness score, the relevance score, and the timeliness score.

6. The computer-implemented method of claim 1 , wherein the encoding scheme encodes an attribute of the insight and an equivalent attribute of the context using the same vector dimension.

7. A computer program product for conversation based data analysis recommendation, the computer program product comprising:

one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions comprising:

program instructions to detect, in a dataset, an insight, the insight comprising a fact about a portion of data in the dataset, the insight maintained in a text-only structured format comprising a plurality of vocabulary terms, each vocabulary term in the plurality of vocabulary terms comprising a field and a plurality of options for values of the field;

program instructions to determine, by analyzing a natural language input, a context, the context identifying an intent and an entity referenced by the natural language input, the context maintained in the text-only structured format;

program instructions to compute, for the insight, a relevance score, the relevance score comprising a measure of similarity between an encoded form of the insight and an encoded form of the context, the encoded form of the insight and the encoded form of the context generated according to the same encoding scheme, the encoded form of the insight comprising an encoding of a vocabulary term used in the insight, the encoded form of the context comprising an encoding of a vocabulary term used in the context, wherein the encoding scheme is a one-hot encoding scheme in which each option in the plurality of options for values of the field is represented by a dimension of a vector; and

program instructions to present a selected insight, the selected insight selected from a set of insights including the insight, the selected insight selected based on the relevance score.

8. The computer program product of claim 7 , further comprising:

program instructions to determine, by analyzing a second natural language input, an updated context;

detecting, responsive to determining that the updated context has below a threshold similarity with the context, a second insight;

program instructions to compute, for the second insight, a second relevance score; and

program instructions to present a second selected insight, the second selected insight selected from a second set of insights including the second insight, the second selected insight selected based on the second relevance score.

9. The computer program product of claim 7 , further comprising:

program instructions to compute, for the insight, an interestingness score;

program instructions to compute, for the insight, a timeliness score; and

program instructions to combine, into an insight score, the interestingness score, the relevance score, and the timeliness score.

10. The computer program product of claim 9 , wherein the selected insight is selected based on the insight score.

11. The computer program product of claim 9 , wherein the insight score comprises a weighted average of the interestingness score, the relevance score, and the timeliness score.

12. The computer program product of claim 7 , wherein the encoding scheme encodes an attribute of the insight and an equivalent attribute of the context using the same vector dimension.

13. The computer program product of claim 7 , wherein the stored program instructions are stored in the at least one of the one or more storage media of a local data processing system, and wherein the stored program instructions are transferred over a network from a remote data processing system.

14. The computer program product of claim 7 , wherein the stored program instructions are stored in the at least one of the one or more storage media of a server data processing system, and wherein the stored program instructions are downloaded over a network to a remote data processing system for use in a computer readable storage device associated with the remote data processing system.

15. The computer program product of claim 7 , wherein the computer program product is provided as a service in a cloud environment.

16. A computer system comprising one or more processors, one or more computer-readable memories, and one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories, the stored program instructions comprising:

program instructions to detect, in a dataset, an insight, the insight comprising a fact about a portion of data in the dataset, the insight maintained in a text-only structured format comprising a plurality of vocabulary terms, each vocabulary term in the plurality of vocabulary terms comprising a field and a plurality of options for values of the field;

program instructions to determine, by analyzing a natural language input, a context, the context identifying an intent and an entity referenced by the natural language input, the context maintained in the text-only structured format;

program instructions to compute, for the insight, a relevance score, the relevance score comprising a measure of similarity between an encoded form of the insight and an encoded form of the context, the encoded form of the insight and the encoded form of the context generated according to the same encoding scheme, the encoded form of the insight comprising an encoding of a vocabulary term used in the insight, the encoded form of the context comprising an encoding of a vocabulary term used in the context, wherein the encoding scheme is a one-hot encoding scheme in which each option in the plurality of options for values of the field is represented by a dimension of a vector; and

program instructions to present a selected insight, the selected insight selected from a set of insights including the insight, the selected insight selected based on the relevance score.

17. The computer system of claim 16 , further comprising:

program instructions to determine, by analyzing a second natural language input, an updated context;

detecting, responsive to determining that the updated context has below a threshold similarity with the context, a second insight;

program instructions to compute, for the second insight, a second relevance score; and

program instructions to present a second selected insight, the second selected insight selected from a second set of insights including the second insight, the second selected insight selected based on the second relevance score.

18. The computer system of claim 16 , further comprising:

program instructions to compute, for the insight, an interestingness score;

program instructions to compute, for the insight, a timeliness score; and

program instructions to combine, into an insight score, the interestingness score, the relevance score, and the timeliness score.

19. The computer system of claim 18 , wherein the selected insight is selected based on the insight score.

20. A data processing environment comprising one or more processors, one or more computer-readable memories, and one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories, the stored program instructions comprising:

program instructions to detect, in a dataset, an insight, the insight comprising a fact about a portion of data in the dataset, the insight maintained in a text-only structured format comprising a plurality of vocabulary terms, each vocabulary term in the plurality of vocabulary terms comprising a field and a plurality of options for values of the field;

program instructions to determine, by analyzing a natural language input, a context, the context identifying an intent and an entity referenced by the natural language input, the context maintained in the text-only structured format;

program instructions to compute, for the insight, a relevance score, the relevance score comprising a measure of similarity between an encoded form of the insight and an encoded form of the context, the encoded form of the insight and the encoded form of the context generated according to the same encoding scheme, the encoded form of the insight comprising an encoding of a vocabulary term used in the insight, the encoded form of the context comprising an encoding of a vocabulary term used in the context, wherein the encoding scheme is a one-hot encoding scheme in which each option in the plurality of options for values of the field is represented by a dimension of a vector; and

program instructions to present a selected insight, the selected insight selected from a set of insights including the insight, the selected insight selected based on the relevance score.

21. A natural language processing system comprising one or more processors, one or more computer-readable memories, and one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories, the stored program instructions comprising:

program instructions to detect, in a dataset, an insight, the insight comprising a fact about a portion of data in the dataset, the insight maintained in a text-only structured format comprising a plurality of vocabulary terms, each vocabulary term in the plurality of vocabulary terms comprising a field and a plurality of options for values of the field;

program instructions to determine, by analyzing a natural language input, a context, the context identifying an intent and an entity referenced by the natural language input, the context maintained in the text-only structured format;

program instructions to compute, for the insight, a relevance score, the relevance score comprising a measure of similarity between an encoded form of the insight and an encoded form of the context, the encoded form of the insight and the encoded form of the context generated according to the same encoding scheme, the encoded form of the insight comprising an encoding of a vocabulary term used in the insight, the encoded form of the context comprising an encoding of a vocabulary term used in the context, wherein the encoding scheme is a one-hot encoding scheme in which each option in the plurality of options for values of the field is represented by a dimension of a vector; and

program instructions to present a selected insight, the selected insight selected from a set of insights including the insight, the selected insight selected based on the relevance score.