Difference between revisions of "Harmonized Data"
(→Calculations) |
(→Missing Values Handling) |
||
Line 25: | Line 25: | ||
==== Missing Values Handling ==== | ==== Missing Values Handling ==== | ||
+ | |||
+ | Values can be "missing" for different reasons, with different implications. dataZoa recognizes several different types of missing values and treats them appropriately. Complete specifications are [[Application_Conventions#Missing_Data_Values|here]]. | ||
===== Calculations ===== | ===== Calculations ===== |
Revision as of 14:10, 5 December 2016
Today's Web is filled with Open Data - hundreds of authoritative, original data sources, freely accessible to everyone.
Before dataZoa, the actual practice of using Open Data sources was time-consuming, messy and painful. A major portion of time was wasted on "data drudgery"; cleaning, aligning, pasting, re-typing and such. All before any real thought or analysis could begin.
As you bring data into dataZoa, it is harmonized (or "normalized") so that any series can work with any other. While we preserve all the original series documentation, we load the dates and values into an idealized time series. This automatically handles the specific nuances of time series data:
Contents
Date Handling
Frequency/Periodicity
http://docwiki.datazoa.com/Application_Conventions#Missing_Data_Values
Date Range
Gaps
Date Formats
Inputs
Outputs
Data Value Handling
Missing Values Handling
Values can be "missing" for different reasons, with different implications. dataZoa recognizes several different types of missing values and treats them appropriately. Complete specifications are here.
Calculations
In all calculations, missing data has the highest precedence; e.g. a number plus an NA yields an NA. Specialized functions in the ComputeCloud can be used for specialized handling, such as carry-forward, etc.
Tables
By default, No Data (ND) formats as a blank, while NA, and NDD data are formatted with "NA" and "NDD", respectively. To fine tune these representations, see Displays.
Charts
Missing values are typically shown as gaps.
Look here for missing value formatting details.