Common Data Anomalies

Data quality problems can come from human errors, system failures or problems during collection and integration. These are the most common ones.

Anomaly	Description	Example	Impact
Missing values	Incomplete records	Blank Age	Lower statistical power, bias
Duplicate records	Repeated entries	Same customer twice	Over-represents observations, skewed results
Incorrect data types	Mismatched formats	"85" stored as text	Calculation errors, analysis failures
Outliers	Extremely high or low values	Age = 900	Distorts statistics and model performance
Inconsistent formats	Different date, case or unit styles	kg and lb mixed	Problems when merging or comparing
Spelling / typographical errors	Errors in text fields	Dhaka, Dhakaa, dhaka	Wrong grouping of categories

Spotting anomalies in a small example

Imagine this small student table:

Row  Name    Date         Score
1    Alice   2026-01-05   82
2    Bob     2026-01-06   75
3    Carol   2026-01-07   90
4    David   2026-01-08   68
5    Eva     2026-01-09   77
6    Alice   2026-01-05   82      <- duplicate of row 1
7    (blank) 2026-01-10   100     <- missing name, possible outlier

Assessing it reveals:

• Rows 1 and 6 are duplicates and may distort the analysis.
• Row 7 has a missing Name, which affects summaries.
• The Date column must stay in one consistent format (YYYY-MM-DD).
• A score of 100 in row 7 may be an outlier, depending on the scoring system.

Each of these has its own fix, covered in the next pages: removing duplicates, filling Name (for example with "Unknown"), standardizing dates, and reviewing the outlier.