Computer systems and methods for automatic generation of models for a dataset

• US 8,860,727 B2
• Filed: 08/16/2011
• Issued: 10/14/2014
• Est. Priority Date: 04/19/2006
• Status: Active Grant

• Current Assignee: Tableau Software Incorporated (Salesforce.com, Inc.)
• Original Assignee: Tableau Software Incorporated (Salesforce.com, Inc.)
• Inventors: Beers, Andrew C., Eldridge, Matthew W., Hanrahan, Patrick M., Taylor, Jonathan E.
• Primary Examiner(s): Vy, Hung T
• Application Number:
  US13/211,048
• Publication Number:
  US 20110302110A1
• Time in Patent Office: 1,155 Days
• Field of Search: 707/899, 707/999.1, 707/798, 345/440, 706/11, 705/35, 705/52
• US Class Current: 345/440
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/283 Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP
• CPC Class Codes: G06F 3/0481 based on specific properties of the displayed interaction object or a metaphor-based environment, e.g. interaction with desktop elements like windows or icons, or assisted by a cursor's changing behaviour or appearance
• CPC Class Codes: G06F 30/00 Computer-aided design [CAD]
• CPC Class Codes: G16B 40/00 ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding

Abstract

A method of automatically generating models from a dataset includes multiple steps. First, a description of a view of a dataset is provided. The description includes multiple fields associated with the dataset. Next, a set of properties is determined for each of the multiple fields. Finally, the description is automatically translated into one or more models based on the respective properties of the multiple fields and a set of predefined heuristics.

Claims

1. A method of automatically generating statistical models for a dataset, comprising:

at a computer system having one or more processors and memory storing programs executed by the one or more processors;

receiving a user description of a graphical view of a dataset, wherein the user description specifies a relationship between a first field and a second field of the dataset;

rendering a graphical visualization of the relationship between the first field and the second field of the dataset using data values associated with the first and second fields in the graphical view of the dataset;

receiving a user instruction to render a statistical model on top of the graphical visualization of the relationship between the first field and the second field of the dataset using data values associated with the first and second fields in the graphical view of the dataset;

in response to the user instruction;

determining a set of data properties for each of the first and second fields; and

determining whether there is a mathematical formula describing the relationship between the first field and the second field of the dataset in accordance with the respective set of data properties; and

when there is a mathematical formula describing the relationship between the first field and the second field of the dataset;

generating one or more statistical models of the dataset in accordance with the mathematical formula and the data values associated with the first and second fields in the dataset; and

rendering at least one graphical representation of the one or more statistical models on top of the graphical visualization of the relationship between the first field and the second field of the dataset in the graphical view of the dataset.

2. The method of claim 1, wherein the first and second fields are partitioned into a plurality of sets, each set associated with a respective region of the graphical view including row, column, page, and an encoding area.

3. The method of claim 1, wherein the data properties of each field include a field type selected from the group consisting of categorical and quantitative and a data type selected from the group consisting of integer, real, Boolean, and time.

4. The method of claim 1, further comprising generating the one or more statistical models in accordance with predefined heuristics for modeling a relationship between the first field and the second field, wherein the predefined heuristics include:

selecting one of the two fields from the user description as an independent variable and the other of the two fields as a response variable of the one or more statistical models.

5. The method of claim 1, wherein each of the one or more statistical models has a model type including linear, polynomial, exponential, logarithmic, and transcendental.

6. The method of claim 1, further comprising rendering the graphical view of the dataset in accordance with the user description.

7. The method of claim 1, wherein generating one or more statistical models further includes comparing the data properties of the first field and the data properties of the second field to determine the one or more statistical models.

8. The method of claim 1, further comprising:

generating a plurality of candidate models using the description, each model having a specific model type and a set of parameters;

automatically selecting from the plurality of models one having a simpler model type and fewer parameters according to predefined criteria.

9. The method of claim 1, wherein the user description is generated by a user dragging and dropping the first and second fields into respective field containers of a graphical user interface.

10. The method of claim 1, wherein the user description is generated by a user selecting the first and second fields as fields of interest through a graphical user interface.

11. The method of claim 1, wherein the user description is generated by a user entering a text string including the first and second fields as fields of interest through a graphical user interface.

12. A computer system, comprising:

memory;

one or more processors; and

at least one program stored in the memory, wherein the at least one program, when executed by the one or more processors, causes the computer system to;

receive a user description of a graphical view of a dataset, wherein the user description specifies a relationship between a first field and a second field of the dataset;

render a graphical visualization of the relationship between the first field and the second field of the dataset using data values associated with the first and second fields in the graphical view of the dataset;

receive a user instruction to render a statistical model on top of the graphical visualization of the relationship between the first field and the second field of the dataset using data values associated with the first and second fields in the graphical view of the dataset;

in response to the user instruction;

determine a set of data properties for each of the first and second fields; and

determine whether there is a mathematical formula describing the relationship between the first field and the second field of the dataset in accordance with the respective set of data properties; and

when there is a mathematical formula describing the relationship between the first field and the second field of the dataset;

generate one or more statistical models of the dataset in accordance with the mathematical formula and the data values associated with the first and second fields in the dataset; and

render at least one graphical representation of the one or more statistical models on top of the graphical visualization of the relationship between the first field and the second field of the dataset in the graphical view of the dataset.

13. The computer system of claim 12, further comprising instructions for generating the one or more statistical models in accordance with predefined heuristics for modeling a relationship between the first field and the second field, wherein the predefined heuristics include:

instructions for selecting one of the two fields from the user description as an independent variable and the other of the two fields as a response variable of the one or more statistical models.

14. The computer system of claim 12, wherein each of the one or more statistical models has a model type including linear, polynomial, exponential, logarithmic, and transcendental.

15. The computer system of claim 12, further comprising instructions for rendering the graphical view of the dataset in accordance with the user description.

16. The computer system of claim 12, further comprising instructions for comparing the data properties of the first field and the data properties of the second field to determine the one or more statistical models.

17. The computer system of claim 12, further comprising:

instructions for generating a plurality of candidate models using the description, each model having a specific model type and a set of parameters;

instructions for automatically selecting from the plurality of models one having a simpler model type and fewer parameters according to predefined criteria.

18. A non-transitory computer readable storage medium storing a computer program product that includes one or more programs configured for execution by a computer system that includes one or more processors and memory, the one or more programs comprising instructions that, if executed by the computer system, cause the computer system to:

receive a user description of a graphical view of a dataset, wherein the user description specifies a relationship between a first field and a second field of the dataset;

render a graphical visualization of the relationship between the first field and the second field of the dataset using data values associated with the first and second fields in the graphical view of the dataset;

receive a user instruction to render a statistical model on top of the graphical visualization of the relationship between the first field and the second field of the dataset using data values associated with the first and second fields in the graphical view of the dataset;

in response to the user instruction;

determine a set of data properties for each of the first and second fields; and

determine whether there is a mathematical formula describing the relationship between the first field and the second field of the dataset in accordance with the respective set of data properties; and

when there is a mathematical formula describing the relationship between the first field and the second field of the dataset;

generate one or more statistical models of the dataset in accordance with the mathematical formula and the data values associated with the first and second fields in the dataset; and

render at least one graphical representation of the one or more statistical models on top of the graphical visualization of the relationship between the first field and the second field of the dataset in the graphical view of the dataset.

19. The computer program product of claim 18, further comprising instructions for generating the one or more statistical models in accordance with predefined heuristics for modeling a relationship between the first field and the second field, wherein the predefined heuristics include:

instructions for selecting one of the two fields from the user description as an independent variable and the other of the two fields as a response variable of the one or more statistical models.

20. The computer program product of claim 18, wherein each of the one or more statistical models has a model type including linear, polynomial, exponential, logarithmic, and transcendental.

21. The computer program product of claim 18, further comprising instructions for rendering the graphical view of the dataset in accordance with the user description.

22. The computer program product of claim 18, further comprising instructions for comparing the data properties of the first field and the data properties of the second field to determine the one or more statistical models.

23. The computer program product of claim 18, further comprising:

instructions for generating a plurality of candidate models using the description, each model having a specific model type and a set of parameters;

instructions for automatically selecting from the plurality of models one having a simpler model type and fewer parameters according to predefined criteria.