Introduction to data science in Artificial Intelligence.pdf

 Emerged from statistics & computer science
 Driven by big data and AI advancements
 Early data analysis was manual , now automated with AI
 Data Science helps AI learn from past data to improve decision
making
2

 AI:
 Focuses on building models that make decisions automatically.
 Uses Machine Learning (ML) and Deep Learning (DL).
 Examples: Chatbots, Image Recognition.
 Data Science:
 Extracts insights from data.
 Uses Statistics,Visualization, and ML.
 Examples:Trend Analysis, Business Forecasting.
3

 Data Science is the process of extracting valuable insights from
data using various techniques.
 It combines Statistics, Data Visualization, and Machine
Learning (ML) to analyze large datasets.
 Helps businesses, healthcare, finance, and AI systems make
data-driven decisions to improve efficiency and accuracy.
 Example 1: Healthcare
 A hospital uses data science to analyze patient records and predict
disease risks based on age, genetics, and lifestyle.This helps doctors
take preventive measures before a disease worsens.
 Example 2: E-Commerce
 Online shopping platforms like Amazon use data science to study
customer preferences and show personalized recommendations,
increasing sales and customer satisfaction.
4

 Automates decision-making (AI models learn from past data)
 Improves accuracy (Reduces errors in manual analysis)
 Enhances efficiency (Faster data processing in AI-driven
systems)
5

 Healthcare: Predicting diseases, AI driven diagnostics
 Finance: Fraud detection, stock market predictions
 Retail: Customer behavior analysis, recommendation systems
 Social Media: Sentiment analysis, content recommendations
6

 Google Search: AI ranks results based on user interactions
 Amazon Recommendations: AI suggests products based on
browsing history
 Netflix: AI personalizes movie suggestions using past viewing
history
 Self-Driving Cars: AI processes real-time sensor data for safe
navigation
7

 Data Collection → Gathering raw data (sensors, APIs, web scraping).
 Data Preprocessing → Cleaning, structuring, removing errors.
 Data Analysis &Visualization → Understanding patterns & insights.
 AI Model Training → Feeding data into ML algorithms.
 AI Model Evaluation → Checking accuracy, avoiding biases, tuning
models.
 AI Decision-Making & Predictions → AI applies learned insights to
make predictions.
 Example:
 Netflix AI learns from past user behavior → Predicts what movies you might
like.
 Self-driving cars process real-time sensor data → Decide when to stop or
accelerate.
8

 Why Ethics Matter in AI?
 AI makes decisions based on data, but if the data is biased, AI can produce
unfair or incorrect results.
 AI is used in hiring, law enforcement, medical diagnosis, and bias in these
systems can cause serious harm.
 Ethical AI ensures fairness, transparency, and accountability in decision-
making.
 What is AI Bias?
 AI bias happens when AI models make unfair or discriminatory decisions
due to biased data.
 If AI is trained on biased historical data, it will continue and reinforce those
biases.
 Example:
 A hiring AI trained on past resumes prefers male candidates because
historically, more men were hired.
 A face recognition AI trained mostly on light-skinned faces struggles to identify
darker-skinned individuals correctly.
9

 What is Data in AI?
 Data is the foundation of AI & Data Science
 AI learns & makes decisions based on processed data. AI decision-
making depends on quality data
 Different AI applications require different types of data.
Understanding different data types helps in better AI models
 Types of Data in AI
 Structured Data: Well organized, stored in tables (e.g., Databases,
Excel)
 Unstructured Data: No predefined format (e.g., Images,Videos,
Social Media Posts)
 Semi Structured Data: Partially organized (e.g., JSON, XML)
10

 Data is properly organized in rows & columns
 Stored in relational databases, spreadsheets, tables
 Easy to search, filter, and analyze
 Used in finance, healthcare, AI training datasets
 Key Features
 Fixed format & follows a defined structure
 Stored in SQL databases
 Easily processable by AI algorithms
 How AI Uses Structured Data?
 AI fraud detection in banking
 AI sales forecasting in businesses
 AI recommendations for users
11

 Examples of Structured Data
 Bank Transactions
 E-Commerce Product Listings
12
Transaction
ID
Customer
Name
Amount Date Status
TXN001 Ali Khan 5000 PKR 10-Jan-2024 Successful
Product ID Name Category Price Stock
101 iPhone 13 Mobile 200,000 PKR In Stock

 Data is partially structured – some organization, but not fully
tabular
 Flexible schema – data format can change
 Stored in NoSQL databases, JSON, XML, CSV
 Key Features
 Combination of structured & unstructured data
 Can include tags, metadata & variable fields
 Used in web applications, emails, social media
 How AI Uses Semi-Structured Data?
 AI spam email filtering
 AI chatbot training
 AI customer behavior analysis
13

 Examples of Semi-Structured Data
 E-Commerce Order JSON Format
14

 Data with no predefined format or structure (e.g.,Text,
Images, Audio,Videos)
 Complex, requires AI to process & analyze
 Stored in Big Data systems & AI models
 Used in Chatbots, Image Recognition, Sentiment Analysis
 Key Features
 Difficult to search & analyze
 Massive in volume (social media, videos, images)
 Requires NLP & deep learning for AI processing
15

 Examples of Unstructured Data
 Social Media Posts & Comments
 "This product is amazing! Highly recommend it!"
 "Worst experience ever. Never buying again!"
 Medical X-Ray & MRI Images
 AI analyzes X-rays to detect diseases
 Used in computer vision & medical diagnostics
 How AI Uses Unstructured Data?
 AI image & face recognition
 AI sentiment analysis of customer reviews
 AI medical diagnosis from scans
16

 Structured Data: Used for AI training, financial predictions
 Unstructured Data: Used for NLP, Image Processing, Chatbots
 AI needs high quality data to learn and make accurate
predictions
 AI models perform better when trained on clean & structured
data
 Semi-structured & unstructured data need advanced AI
techniques
 Data Science helps convert raw data into useful insights
17

 Netflix collects user viewing history and preferences.
 Uses AI models to suggest personalized recommendations.
 Data Science ensures continuous improvement in suggestions.
18

Introduction to data science in Artificial Intelligence.pdf

More Related Content

Similar to Introduction to data science in Artificial Intelligence.pdf (20)

Recently uploaded (20)

Introduction to data science in Artificial Intelligence.pdf