41. What is data cleaning?
Data cleaning is the process of fixing or removing incorrect, incomplete, or inconsistent data.
Example: Removing duplicate customer records, Fixing wrong date formats.
42. How do you handle missing data?
Common methods:
- Remove rows (if few missing)
- Replace with mean, median, or 0
- Use forward or backward fill
Example (SQL):
SELECT COALESCE(sales, 0) AS sales FROM orders;43. How do you treat outliers?
- Identify using sorting, box plots, or Z-score
- Remove or cap extreme values
Example: Sales = 10,000, 12,000, 15,000, 1,00,000 → outlier.
44. What is data normalization?
Scaling data between 0 and 1.
Example: Normalized value = (x - min) / (max - min)
Used in ML and comparisons.
45. What is data standardization?
Centers data around mean 0 with std dev 1.
Example: Z = (x - mean) / std
46. How do you check data quality?
- Accuracy
- Completeness
- Consistency
- Validity
- Timeliness
Example: Sales should never be negative.
47. What is duplicate data?
Same record appearing more than once.
Example: Same customer ID repeated multiple times.
48. How do you validate source data?
- Compare with source systems
- Check row counts
- Verify key metrics
Example: Total revenue in report = total revenue in database.
49. What is data transformation?
Converting data into usable format.
Examples:
- Converting dates
- Creating new columns
- Aggregating values
50. Why is data preparation important?
Clean data = correct insights. Poor data leads to wrong decisions.
Example: Wrong sales data → wrong business strategy.
Double Tap ♥️ For Part-6