Normalization

Normalization means changing data into a consistent format or measurement basis. The word has several meanings, so a useful explanation states exactly what changed: dates, units, category labels, or the scale used for comparison. Saying that a dataset is “normalized” does not describe the method by itself.

What consistency makes possible

Records from different sources may express the same kind of information differently. One file may store a duration in seconds and another in minutes. Dates may use different orders or time zones. These differences need to be resolved before values are combined or compared.

The U.S. Geological Survey explains that standards must address both format and meaning. It also recommends retaining the original value when standardization would lose information. Those principles are useful when preparing public-safety records for analysis. USGS data standards.

Example

Hypothetical: Agency A reports a response interval as 360 seconds. Agency B reports an interval as six minutes. Converting 360 seconds to six minutes puts both values in the same unit.

That conversion does not establish that the intervals are comparable. Agency A might measure from call receipt while Agency B starts at dispatch. Matching units puts values on the same measurement scale; matching start and end points establishes what the interval measures. Both are needed before describing the two responses as equally long.

Common confusion

In a statistical comparison, normalization may mean expressing a count relative to a population or another base. A crime rate per 100,000 residents is one example. The method must identify the count, population, period, and multiplier.

In database design, normalization has a separate technical meaning involving how information is organized across tables. A reader should not have to guess which meaning a dashboard uses.

Standardization also does not correct every source error. Converting a mistyped duration into another unit preserves the error in a new form. And assigning two local call labels to a shared category requires an explanation of why their meanings are compatible.

Reading the method

A clear note names the input, output, and rule: “Durations are converted from seconds to minutes,” for example. For dates, it identifies the time zone. For combined categories, it shows which original labels belong to each group.

Preserving the original fields makes the transformation easier to check. It also helps readers distinguish an agency's own terminology from a standardized label added later for comparison.

Related terms

• Data dictionary
• Crime rate
• Aggregation
• Data validation

Sources

• U.S. Geological Survey: Data Standards — Supports consistent formats and meanings, unit-level standards, documentation, and preservation of original detail.