DATA PREPARATION
Data quality: fix the data before it feeds a decision
Data quality work with TIMi uses Anatella to detect duplicates, inconsistent values and spelling errors, then clean them with rules you can read, including fuzzy matching for names and addresses.
Detect
Duplicates, missing values and inconsistent formats.
Correct
Spelling errors in names and address fields.
Match
Fuzzy joins between tables that do not share a clean key.
HOW IT WORKS
How data quality work happens.
Three steps, in one engine, on a standard server with no cluster.
1. Profile the data
Anatella extracts meta-data and checks keys, for example by listing duplicate keys before a join.
2. Clean and standardize
Spelling correction, format standardization and fuzzy matching bring values to a common form.
3. Verify in the flow
Checks are steps of the pipeline, so they run every time the data is refreshed.
GETTING IT RIGHT
Choices that shape a cleaning project.
Cleaning rules encode business decisions. They need an owner.
Decide what counts as an error
Some irregular values are mistakes and others are real. The business owner of the data decides which is which.
Keep the raw data
Cleaning is done in a flow that produces a new table, so the original data stays available for audit.
Measure before and after
Counts of corrected, rejected and unchanged rows show the effect of the rules and help refine them.
TRANSPARENCY
Cleaning rules you can read.
Each correction is a step in a graph. Data owners and auditors can see which rules were applied, in which order, and review them without reading code.
For data owners
See which rules change which values.
For audit
The original data is kept, and the cleaning is documented in the flow.
For operations
The same rules apply at every refresh.
DATA
What data can be cleaned.
Anatella connects to more than 80 sources, so this data can be gathered without prior custom development.
Names and addresses
Spelling correction and fuzzy matching.
Reference tables
Duplicate keys and inconsistent codes.
Free text
Text fields standardized before analysis.
Source extracts
Data from any of the connected sources.
PROOF
Working on text fields.
We do not have a published case study dedicated to data quality. The reference below is indirect: it relies on text processing, one of the cleaning techniques.
Text mining refreshed in under 6 hours
Not documented as a data quality case. It shows text processing running on a regular refresh cycle of under six hours.
FAQ
Common questions.
Can it fix addresses and names automatically?
It can correct spelling errors and standardize values, and match records fuzzily. Results should be reviewed on a sample before being applied to the whole dataset.
Does cleaning change my source data?
No. The flow writes a new cleaned table. The source stays as it is.
Is there a limit to the size of the data?
No fixed limit. Anatella cleans large tables on a standard server, and the practical limit is the hardware.
Who defines the cleaning rules?
Your data owners. TIMi provides the operators, and your teams decide which values count as errors.
RELATED USE CASES
Keep exploring.
SEE FOR YOURSELF
Test it on your own data.
Run TIMi on your own data with a free trial, or talk to us about your case.
