US 12,169,503 B2
Visual data computing platform using a progressive computation engine
Zeyuan Shang, Cambridge, MA (US); Emanuel Zgraggen, Malden, MA (US); Tim Kraska, Arlington, MA (US); Benedetto Buratti, Providence, RI (US); Philipp Eichman, Melrose, MA (US); Navid Karimeddiny, Boston, MA (US); Charles Meyer, Cambridge, MA (US); and Wesley Runnels, Cambridge, MA (US)
Assigned to Einblick Analytics, Inc., Cambridge, MA (US)
Filed by Einblick Analytics, Inc., Cambridge, MA (US)
Filed on May 15, 2023, as Appl. No. 18/197,598.
Application 18/197,598 is a continuation of application No. 17/869,243, filed on Jul. 20, 2022, granted, now 11,651,007, issued on May 16, 2023.
Claims priority of provisional application 63/224,002, filed on Jul. 21, 2021.
Prior Publication US 2023/0289364 A1, Sep. 14, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 16/26 (2019.01); G06F 16/16 (2019.01); G06F 16/23 (2019.01)
CPC G06F 16/26 (2019.01) [G06F 16/168 (2019.01); G06F 16/2379 (2019.01)] 10 Claims
OG exemplary drawing
 
1. A computer program product in a computer-readable medium, the computer program product comprising program code that, when executed by a processor, is configured to:
render a visual workspace with a set of display elements, the set of display elements including one of: an operator associated with a block of computation, and a dataframe that is a structured or semi-structured piece of data generated from a data source or an operator;
responsive to receipt of data indicating a change to an operator or a dataframe, receive a progressive data stream of responses from a processing engine that is configured as an accelerator between the visual workspace and a dataset comprising a plurality of data sources, the progressive data stream of responses representing a computation by the processing engine over data stored in one or more of the plurality of data sources and comprising a first response that is an approximation, one or more incremental updates, and an optional final response, and wherein the first response is returned based on an initial subset or sample of the dataset, and wherein as the computation by the processing engine iterates by scaling over the dataset, results are progressively refined; and
update a state of the visual workspace dynamically using the progressive data stream of responses.