Intro to data science sumary

What is Data Science?

An interdisciplinary field combining math, statistics, computer science, and domain knowledge to extract insights from data.

What’s the difference between AI, ML, and Data Science?

• AI = Machines performing tasks requiring human-like intelligence.

• ML = Subset of AI that learns patterns from data.

• Data Science = Field that prepares and analyzes data used to train ML/AI.

What are the 5 components of the Data Science process?

Data Collection, Cleaning, Analysis, Visualization, Decision Making.

What are the 7 Vs of Big Data?

Volume, Velocity, Variety, Veracity, Value, Variability, Visualization.

Give an example of each data type:

• Structured: SQL table, Excel sheet.

• Semi-structured: JSON, XML.

• Unstructured: Image, video, audio, free-text.

Why is Data Quality important?

High-quality data builds trust, improves decision making, and prevents misleading results.

Name 3 data cleaning techniques for structured data.

Remove duplicates, handle missing values, fix incorrect entries.

What’s “imputation”?

Filling missing data with estimated values (e.g., mean, median).

Name 2 techniques for cleaning unstructured data.

Text cleaning (remove stop words, fix spelling), image preprocessing (resizing, normalization).

What are epochs, batch size, and learning rate?

• Epoch: One full pass through the dataset during training.

• Batch size: Number of samples per training step.

• Learning rate: How much the model’s weights change per update.

What’s the difference between training, validation, and test data?

• Training set: Teaches the model.

• Validation set: Tunes model & prevents overfitting.

• Test set: Evaluates final model performance.

Image Recognition vs Object Detection?

• Recognition: Identifies what is in an image.

• Detection: Identifies what and where by drawing bounding boxes.

What are Precision and Recall?

• Precision: Correct positive predictions / all positive predictions.

• Recall: Correct positive predictions / all actual positives.

What is F1 Score?

Harmonic mean of Precision and Recall → balances both.

What is mAP@50-95?

Mean Average Precision measured across IoU thresholds 0.5–0.95; evaluates detection accuracy.

What does IoU stand for and what does it measure?

Intersection over Union — overlap between predicted and actual bounding boxes.