Generalizing spreadsheet computation for evolving spreadsheets at scale

Wattanawaroon, Tana

Permalink

https://hdl.handle.net/2142/113887 Copy

Description

Title

Generalizing spreadsheet computation for evolving spreadsheets at scale

Author(s)

Wattanawaroon, Tana

Issue Date

2021-12-02

Director of Research (if dissertation) or Advisor (if thesis)

Parameswaran, Aditya

Doctoral Committee Chair(s)

Parameswaran, Aditya

Committee Member(s)

• Chekuri, Chandra
• Karahalios, Karrie
• Roy, Sudeepa

Department of Study

Computer Science

Discipline

Computer Science

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• spreadsheets
• databases
• scale
• interactivity
• versioning

Abstract

"Spreadsheets are one of the most ubiquitous ad-hoc data analysis and manipulation tools. Their strength over traditional relational database management systems lies in their ability to allow users to manipulate data interactively through an intuitive interface. However, the capabilities of current spreadsheet systems to handle datasets that evolve over time are limited in several dimensions: (a) limited power: it is difficult to perform relational-style queries, which is often needed for large data analysis, while keeping the convenience of formula-like automatic recalculation, (b) limited introspection: the ability to reason about the source of changes between versions at a higher level is often unsupported, and (c) limited interactivity: the computation in spreadsheets at scale can make the system unresponsive, rendering the strength of spreadsheets moot, (d) limited structure utilization: the computation in spreadsheets often fails to utilize the semi-structured nature of real-world spreadsheets. The dissertation discusses developments that overcome these hurdles. First, we discuss an extension to spreadsheet formulae that allows for relational-style queries in a manner that is consistent with typical formula computation engines. Second, we develop the theory of ""diffing"", representing data updates in a concise manner. Third, we introduce Asynchronous Formula Computation, a technique that improves spreadsheet interactivity when dealing with formula computation, while guaranteeing consistency of the results. Finally, we improve formula computation by utilizing structures of real-world spreadsheets and building a more concise representation."

Graduation Semester

2021-12

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/113887

Copyright and License Information

Copyright 2021 Tana Wattanawaroon

Owning Collections