What is Outliers in Data? | Simple Explanation for Class 12

S7-SA3-0229

What is the Concept of Outliers in Data?

Grade Level:

Class 12

AI/ML, Physics, Biotechnology, FinTech, EVs, Space Technology, Climate Science, Blockchain, Medicine, Engineering, Law, Economics

Definition

What is it?

Outliers are data points that are very different from most other data points in a dataset. They lie an unusual distance from the rest of the data, either much higher or much lower. Think of them as the 'odd ones out' in a group of numbers.

Simple Example

Quick Example

Imagine your class had a science test. Most students scored between 60 and 80 marks. But one student scored 5 marks, and another scored 98 marks. These 5 and 98 marks are likely outliers because they are very far from what most students scored.

Worked Example

Step-by-Step

Let's find outliers in a set of daily temperatures in degrees Celsius: 28, 30, 29, 31, 27, 45, 29, 30.

Step 1: Arrange the data in ascending order: 27, 28, 29, 29, 30, 30, 31, 45.
---Step 2: Find the First Quartile (Q1) and Third Quartile (Q3). Q1 is the median of the first half, Q3 is the median of the second half. For our 8 data points, Q1 is the average of the 2nd and 3rd values (28 and 29), so Q1 = (28+29)/2 = 28.5. Q3 is the average of the 6th and 7th values (30 and 31), so Q3 = (30+31)/2 = 30.5.
---Step 3: Calculate the Interquartile Range (IQR). IQR = Q3 - Q1 = 30.5 - 28.5 = 2.
---Step 4: Calculate the lower bound for outliers: Q1 - 1.5 * IQR = 28.5 - 1.5 * 2 = 28.5 - 3 = 25.5.
---Step 5: Calculate the upper bound for outliers: Q3 + 1.5 * IQR = 30.5 + 1.5 * 2 = 30.5 + 3 = 33.5.
---Step 6: Identify data points outside these bounds. Any value less than 25.5 or greater than 33.5 is an outlier. Our data includes 45, which is greater than 33.5.

Answer: The outlier in this dataset is 45.

Why It Matters

Understanding outliers is super important because they can mislead our analysis in fields like AI/ML, FinTech, and Climate Science. For example, in medicine, an outlier in patient data could signal a rare disease or an error. Spotting them helps engineers build safer EVs and enables scientists to make better predictions about climate change.

Common Mistakes

MISTAKE: Assuming all unusual data points are errors. | CORRECTION: Outliers can be errors, but they can also represent genuine, rare events or important new information. Always investigate them before removing.

MISTAKE: Only looking for very high values as outliers. | CORRECTION: Outliers can be extremely low values too. A data point much smaller than the rest is just as much an outlier as one much larger.

MISTAKE: Removing outliers without proper justification. | CORRECTION: Removing outliers can sometimes hide important patterns or problems. Only remove them if you have a strong reason, like knowing it was a data entry mistake.

Practice Questions

Try It Yourself

QUESTION: In a cricket match, a player scored runs: 5, 12, 8, 15, 100, 7, 10. Is there an obvious outlier? | ANSWER: Yes, 100 is an obvious outlier because it is much higher than the other scores.

QUESTION: A class collected data on daily mobile data usage (in GB): 1.2, 1.5, 0.8, 1.3, 1.1, 8.0, 1.0. Using the 1.5*IQR rule (Q1=1.0, Q3=1.3), identify the outlier. | ANSWER: IQR = 1.3 - 1.0 = 0.3. Lower bound: 1.0 - 1.5*0.3 = 0.55. Upper bound: 1.3 + 1.5*0.3 = 1.75. The value 8.0 is greater than 1.75, so 8.0 is the outlier.

QUESTION: A factory measures the weight of 10 biscuit packets (in grams): 98, 102, 99, 101, 100, 103, 100, 105, 100, 10. Identify the outlier(s) using the IQR method. (Hint: Sorted data: 10, 98, 99, 100, 100, 100, 101, 102, 103, 105) | ANSWER: Q1 = 99, Q3 = 102. IQR = 3. Lower bound = 99 - 1.5*3 = 94.5. Upper bound = 102 + 1.5*3 = 106.5. The value 10 is less than 94.5, so 10 is the outlier.

MCQ

Quick Quiz

Which of the following is most likely an outlier in a dataset of daily auto-rickshaw fares (in Rupees) for a 5 km ride?

800

The Correct Answer Is:

Most auto-rickshaw fares for a 5 km ride would be in the range of 60-100 Rupees. A fare of 800 Rupees is extremely high compared to the others, making it a clear outlier.

Real World Connection

In the Real World

In FinTech, banks use outlier detection to spot fraudulent transactions. If someone usually spends 500-1000 rupees on UPI, but suddenly there's a 50,000 rupee transaction, the system flags it as a potential outlier. This helps protect customers from scams.

Key Vocabulary

Key Terms

DATASET: A collection of related information or numbers. | MEDIAN: The middle value in a sorted list of numbers. | QUARTILE: Values that divide a dataset into four equal parts. | IQR (INTERQUARTILE RANGE): The range between the first and third quartiles, showing the spread of the middle 50% of the data. | DATA POINT: A single piece of information or value in a dataset.

What's Next

What to Learn Next

Great job understanding outliers! Next, you should explore 'How to Handle Outliers'. Knowing how to identify them is just the first step; learning what to do with them is crucial for accurate data analysis in any field.