Time Series Data Aggregation Using Floor and Least Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional time series data aggregation methods, such as M4, face challenges in the IoT context due to uneven data distribution, inaccurate division of timestamp-based data, and lack of consideration for data types and AI/predictive analytics, leading to visualization errors and inefficiencies.
Innovation Solution
An improved querying method that includes a floor function and least function to ensure accurate division of data into the correct number of groups, filters out redundant values, and applies AI/predictive analytics for data interpolation when no data exists in a requested time range, ensuring accurate and efficient time series visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional aggregation methods (M4) are used to reduce data volume, then data reduction rate is improved, but visualization accuracy deteriorates due to incorrect division of timestamp-based data into N+1 groups instead of N groups
Solution Approach 1:
The patent changes the grouping parameter calculation method by introducing floor() and least() functions to correctly compute group assignments. This ensures data is divided into exactly N groups as requested, rather than N+1 groups, thereby maintaining visualization accuracy while achieving data reduction.
Solution Approach 2:
The patent replaces traditional rounding-based grouping mechanics with a floor-based grouping mechanism. This substitution eliminates the off-by-one error inherent in rounding methods and ensures precise control over the number of groups, resolving the contradiction between data reduction and accuracy.
2Productivity
If traditional aggregation methods are used, then processing speed is improved, but visualization reliability deteriorates due to inclusion of redundant values and lack of data type consideration
Solution Approach 1:
The patent extracts and removes redundant values from each group, keeping only the minimum and maximum values with their associated timestamps. This extraction process eliminates duplicate information that would otherwise cause visualization errors, improving reliability without significantly impacting processing speed.
Solution Approach 2:
The patent applies different processing rules to different data types (numeric vs. non-numeric) and different value positions (min vs. max). This localized quality approach ensures that each data point is handled according to its specific characteristics, enhancing visualization reliability while maintaining efficient processing.
3Quantity of substance
If data is divided into N groups using traditional methods, then data reduction is achieved, but measurement accuracy deteriorates due to incorrect timestamp assignment and redundant values
Solution Approach 1:
The patent replaces traditional rounding-based group assignment with floor-based group assignment. This mechanical substitution ensures that timestamps are correctly assigned to groups without the off-by-one error, maintaining timestamp accuracy while achieving the desired data reduction to N groups.
Solution Approach 2:
The patent performs preliminary calculations of group boundaries and assignments before actual data aggregation. This preliminary action ensures that each data point is correctly assigned to its designated group from the outset, preventing timestamp assignment errors and redundant value inclusion.
Data Source
AI summary
Methods, systems, and computer-readable storage media for receiving a query that is coded into a computer-executed application that queries a database system, the query including a first portion that defines a number of groups that data is to be divided into, and a second portion that removes redundant values from a group, if any, processing, within the database system, the query to perform a plurality of computations within the database system, and produce a result set including a plurality of data groups, each data group having a minimum value and associated timestamp, and a maximum value and associated timestamp, and transmitting the result set to the application to provide one or more time series visualizations for display in a user interface.