AI-417-IX Unit 2 Data Literacy Session 2.pdf
Transcript
Unit 2 Data Literacy Session-2 Acquiring Data, Processing, and Interpreting Data Learning Outcomes · Familiarizing youth with different data terminologies like data acquisition, processing, analysis, presentation, and interpretation. · Discussing diffe...
Unit 2 Data Literacy Session-2 Acquiring Data, Processing, and Interpreting Data Learning Outcomes · Familiarizing youth with different data terminologies like data acquisition, processing, analysis, presentation, and interpretation. · Discussing different methods of data interpretation like qualitative and quantitative. · Understanding the methods and different collection techniques. · Critically think about their advantages and disadvantages. · Identifying various data presentation methods with examples and interpreting them. · Gain awareness about the advantages and impact of Data interpretation on business growth. Activity Top-Secret Weather Mission: Find the Perfect Dataset! Calling all AI Agents! Today, we're on a mission to train a weather-predicting AI model. But first, we need the fuel: data! Mission Briefing: · You'll be working in pairs, like a crack detective team. · Your mission: Locate a top-secret online dataset containing weather forecast information. · This data will be used to train our AI to predict future weather patterns, becoming a weather-reading whiz! Operation: Dataset Discovery Here are your intel sources (websites) to infiltrate: 1. Kaggle: A treasure trove of datasets, hidden in plain sight! (Search for "weather forecast dataset") 2. Quandl: Another intel hub, brimming with financial and economic data, but it might also hold weather secrets! (Search for "weather data") 3. Open Weather Map: This website might have its own weather data stashed away. Let's see what intel they offer! (Search for "historical weather data") Mission Objective · Infiltrate each website and search for weather forecast datasets. · Capture Intel: Once you find a promising dataset, take a screenshot or copy an image of it. · Mission Report: After visiting all three websites, gather with your partner and compare your findings. Discuss: 1 o Which dataset seems most relevant for training our AI weather predictor? o What kind of weather data does each dataset contain? (Temperature, precipitation, wind speed, etc.) o Are there any limitations or restrictions mentioned for using the data? Mission Success By working together and sharing your intel, you'll be well on your way to finding the perfect dataset to train your AI weather model! Remember, teamwork makes the dream work (and the weather forecast more accurate)! 1.1 Types of Data in AI 1. Textual Data (Qualitative Data) o Composed of words and phrases. o Used in Natural Language Processing (NLP) tasks like sentiment analysis or text summarization. o Example: Search queries ("Which is a good park nearby?"). 2. Numeric Data (Quantitative Data) o Consists of numbers. o Used for statistical analysis and modelling. o Examples: Cricket score, restaurant bill amount. Further Classification of Numeric Data · Continuous Data: Numeric data with a continuous range of values. Examples: height, weight, temperature, voltage. · Discrete Data: Numeric data consisting only of whole numbers and cannot be fractional. Example: number of students in a class. Types of Data used in three domains of AI · Computer Vision (CV): Visual Data: Images, Videos. · Natural Language Processing: Textual Data: Documents, PDF files. · Statistical Data (SD): Numeric Data: Tables, Excel Sheets CV · Visual Dat e.g Images, Videos NLP · Tzxtual Data e.g. Documents, PDF files SD · Statistical Data e.g. Tables, Excel Sheets 2 Visual Data e.g. Image, Videos Textual Data e.g. Documents, PDF files Numeric Data e.g Tables, Excel Sheets Pick & Choose (Quantitative or Qualitative?) Temperature Gender Shoe Size Favourite Colour Body Weight 1.2 Data Acquisition / Acquiring Data Data Acquisition, also known as acquiring data, refers to the procedure of gathering data. This involves searching for datasets suitable for training AI models. The process typically comprises three key steps: Data is the fuel that powers Artificial Intelligence (AI). But what exactly is data, and how does it get used to create intelligent machines? This interactive guide will take you on a journey to explore: 3 · Where Data Comes From: Dive into different types of data, how we find it, and even how to create new data! · Making Data Usable: Discover the secrets behind cleaning and organizing data to make it ready for AI. · Speaking Data's Language: Learn about features, the unique characteristics that define each piece of data. · Seeing the Bigger Picture: Explore how data is interpreted and visualized to reveal hidden patterns and insights. Let's Get Started: Data Acquisition - The Data Hunt! Imagine you're building a self-driving car. To teach it to navigate the road, you'll need data! What kind of data would be useful? To navigate the roads safely, your car needs to "see" its surroundings. So, what kind of data would be useful? we want to collect data for making a CV model for a self-driving car. · Data Discovery: This is like a treasure hunt for data! We can search online databases or even capture videos while driving to collect: o Images: Pictures of roads, traffic signs, pedestrians, and other objects. o Sensor Data: Information from LiDAR, radar, and cameras to measure distances and surroundings. 4 Activity Data Detective 1. List three types of data you'd need for your self-driving car (e.g., pictures of roads, traffic signs). 2. How could you find this data? (e.g., search online databases, record videos while driving). Sample Data Augmentation · Data augmentation means increasing the amount of data by adding copies of existing data with small changes · The image given here does not change, but we get data on the image by changing different parameters like color and brightness · New data is added by slightly changing the existing data. Acquiring Data – Sample Data Generation · Data generation refers to generating or recording data using sensors. · Recording temperature readings of a building is an example of data generation. · Recorded data is stored in a computer in a suitable form Sources of Data: Various Sources for Acquiring Data · Primary Data Sources — Some of the sources for primary data include surveys, interviews, experiments, etc. The data generated from the experiment is an example of primary data. Here is an excel sheet showing the data collected for students of a class. 5 · Secondary Data Sources - Secondary data collection obtains information from external sources, rather than generating it personally. Some sources for secondary data collection include: Kaggle is an online community of data scientists where you can access different types of data. Countries like India, Australia, New Zealand, & Singapore are openly sharing datasets on various parts Google dataset search is a toolbox by google that can search for data by name. UCI is a collection of databases, domain theories, and data generators in collaboration with the University of Massachusetts. 1.3 Best Practices for Acquiring Data Checklist of factors that make data good or bad Good Data Bad Data Information is well structured Inofmration is scattered It is accurate Contains lots of incorrect values It is consistent Contains missing and duplicate values It is cleanly presented It is poorly presented Contains information which is relevant to Contains information which is not our requirement relevant to our requirement Example: Imagine a dataset with student names and test scores. Some names might be misspelled, and some scores might be missing. Data cleaning would fix these issues to ensure accurate analysis. 6 Activity Activity 2: Data Doctor Look at the data table below (containing student names and grades, with some errors). Identify and correct two data cleaning issues: Student Name Grade Kalpana 95 Harish 102 (highest possible score is 100) Rahul N/A Data acquisition from websites Data acquisition from websites, also known as web scraping, can be a valuable tool for gathering information for AI projects. However, it's important to approach it ethically and responsibly ETHICAL CONCERNS IN DATA ACQUISITION Data acquisition, the foundation of AI, comes with ethical considerations. Bias Take steps to understand and avoid any Let's explore some potential pitfalls preferences or partiality in data and how to navigate them proactively: Take necessary permissions before Consent Data Dilemmas: Keeping Things Fair collecting or using an individual's data for Our AI Friends! Transparency Explain how you intend to use the Imagine you're training a robot dog to collected data and do not hide intentions fetch! You show it pictures of balls... mostly red ones. What happens when it Anonymity Protect the identity of the person who is sees a blue ball? Uh oh, confusion! This the source of data is kind of like what can happen with data. Let's explore some things to keep Accountability Take responsibility for your actions in case in mind when collecting data for AI of misuse of data projects, like our robot dog trainer. 7 1. Fairness First: Avoiding Bias Imagine: A robot judge uses data to decide court cases. But what if the data mostly shows people with a certain hair color getting harsher punishments? That's not fair! We need: Data that represents everyone, not just a certain group. This helps AI avoid making biased decisions. 2. Permission is Key: Getting Consent Imagine: You borrow your friend's phone to play a game, but it secretly collects all their messages! Not cool, right? We need: People to know and agree to their data being used for AI projects. It's like asking permission before borrowing something! 3. Openness is Best: Transparency Imagine: Your robot dog trainer hides how it teaches the dog to fetch. Confusing, right? You don't know if it's using the best methods! We need: To understand how AI works and why it makes certain decisions. This helps us trust it and fix any problems. 4. Keeping Secrets: Anonymity Imagine: The robot judge uses data with people's names and addresses! This could be a privacy nightmare! We need: To keep people's identities hidden when using their data for AI. It's like protecting your secret handshake with your friend! 5. Taking Responsibility: Accountability Imagine: The robot judge makes a mistake, but no one knows why! How can we fix it? We need: To be able to track who is responsible for how AI is developed and used. That way, we can fix any problems that arise. Proactive Measures Here are some steps you can take to address these concerns proactively: · Impact Assessment: Conduct an impact assessment to identify potential ethical risks associated with data collection and AI development. · Data Governance Framework: Establish a data governance framework that outlines data collection practices, access controls, and security measures. · Ethical AI Principles: Adopt ethical AI principles that guide your data acquisition and AI development processes. Benefits of Ethical Data Acquisition By proactively addressing ethical concerns, you can reap several benefits: · Builds Trust: Demonstrating ethical data practices builds trust with users and stakeholders. · Reduces Risks: Mitigates potential legal and reputational risks associated with unethical data practices. · Enhances Outcomes: Leads to fairer, more reliable, and socially responsible AI models. 8 1.4 Features of Data and Data Pre-processing Usability of Data There are three primary factors determining the usability of data: Data Wrangling: From Messy to Meaningful Data often comes in a raw, unorganized state. This is where data processing steps in! · Cleaning: Removing duplicates, fixing errors, and filling in missing information. · Augmentation: Creating new data variations from existing data (e.g., rotating images). Usability of Data: Making it Work Just like a messy room needs cleaning before you can play, data often needs some work before it's truly useful. This process is called data pre-processing, and it's like organizing your room to make it easier to find things. Here are three key factors that affect how usable data is: · Structure: Imagine clothes thrown all over the floor. Structured data is like putting clothes in drawers and organizing them neatly (e.g., shirts in one drawer, pants in another). This makes it easier to find what you need. Another example of the structure can be understood by the following · Cleanliness: Data can be messy too! Clean data is free from errors, like missing information or duplicates. Imagine finding two identical shirts in the same drawer (duplicates). We would remove one to keep things tidy. Missing socks (missing data) might need to be replaced or the sock might be removed from the drawer entirely. In this particular example, duplicate values are removed after cleaning the data. 9 · Accuracy: Think of a ruler that measures incorrectly. Accurate data reflects real-world values. Imagine measuring the length of a book with a ruler and getting 20 centimetres. We would check this measurement against the ruler to ensure it's accurate. In this particular example, we are comparing data gathered from measuring the length of a small box in centimetres. More Examples · Duplicate Data: Imagine collecting student names, but some names are spelled twice (e.g., Rohit Kumar, Rohit Kumar). Data cleaning removes these duplicates. · Missing Values: Maybe some students' test scores are missing. Data cleaning might estimate those scores or remove those entries if they're not important. · Accuracy Check: We might measure the height of a plant and get 50 centimetres. Data cleaning compares this data (50 cm) to the actual measurement to ensure it's accurate. Test your knowledge A. What kind of data is more usable, according to you? ________________________________________________________________________________________________ ________________________________________________________________________________________________ ________________________________________________________________________________________________ B. If we have a lot of data which is not clean, is it good for AI? ________________________________________________________________________________________________ ________________________________________________________________________________________________ ________________________________________________________________________________________________ Data Features: Building Blocks of Information · Definition: Data features are the individual characteristics that describe each piece of data within a dataset. They act like building blocks, providing details about the information. Or Data features are the characteristics or properties of the data. They describe each piece of information in a dataset. Examples: o In a student record dataset, features might be name, age, grade, or attendance. o In an image dataset, features could be color, brightness, or object shapes. 10 · Importance: Features allow us to understand and analyze data. By examining them, we can identify patterns, trends, and relationships within the information. Independent and Dependent Features: Powering AI Models · Two Types: In the context of AI models, there are two crucial feature categories: o Independent Features (Input): These are the features we provide to the model. They act as the starting point for the model's analysis and are used to make predictions. o Dependent Features (Output): These are the results generated by the model. They represent what the model is trying to predict based on the independent features. Analogy: Imagine you're feeding data to a fortune teller (the AI model). · Independent Features: You tell the fortune teller your palm lines and birthdate (the input). · Dependent Feature: The fortune teller predicts your future (the output). 11 3. Data Processing and Data Interpretation Data processing and interpretation have become very important in today's world. Can you answer this? · Niki has 7 candies, and Ruchi has 4 candies · How many candies do Niki and Ruchi have in total? o We can answer this question using data processing · Who should get more candies so that both Niki and Ruchi have an equal number of candies? · How many candies should they get? o We can answer this question using data interpretation Data Processing: Cleaning Up the Mess Imagine you have a box filled with random receipts, photos, and notes. This is like raw data - it's full of information, but it's difficult to understand. Data processing is like sorting through that box. You might: · Organize: Group similar items together (categorize data) · Clean: Remove irrelevant information (fix errors) · Format: Convert data into a format which computer can understand (numbers, codes) By processing the data, you make it usable for the next step... Data Interpretation: Telling the Story Now you have a neat pile of receipts, categorized photos, and clear notes. Data interpretation is about making sense of it all. You might: 12 · Analyze: Look for patterns and trends (identify relationships) · Draw conclusions: Explain what the data means (answer questions) · Communicate: Share your insights with others (present findings) 4. Understanding some keywords related to Data The Data Journey: From Collection to Insight Data is all around us, but it takes work to turn it into actionable knowledge. Here's a breakdown of the key steps in this journey: 1. Acquisition: Gathering the Raw Ingredients This is where you collect data from various sources. It could be: · Surveys: Asking people questions directly. · Sensors: Gathering data from devices like fitness trackers or weather stations. · Databases: Existing collections of information. · Websites and social media: Scraping public data online (with permission!). 2. Processing: Cleaning and Organizing the Mess Imagine raw data as a messy kitchen after a big cookout. Processing is like cleaning up: · Sorting: Organizing data into categories based on its type (numbers, text, etc.). · Cleaning: Fixing errors and removing irrelevant information. · Formatting: Converting data into a format that computers can understand (numbers, codes). 3. Analysis: Sifting Through the Ingredients Now that your data is organized, it's time to analyze it. This involves: · Identifying patterns: Looking for trends and relationships within the data. · Statistical analysis: Using calculations to understand the data's meaning (averages, medians). 4. Interpretation: Making Sense of the Dish Data interpretation is like explaining the flavor profile of your culinary creation. Here, you: · Draw conclusions: Based on the analysis, explain what the data means. · Answer questions: Use the data to answer the questions you set out to answer. 13 5. Presentation: Sharing the Feast The final step is to share your findings with others. This involves: · Visualization: Creating charts, graphs, or images to present the data clearly. · Reporting: Writing a report that explains your findings and conclusions. Methods of Data Interpretation How to interpret Data? Based on the two types of data, there are two ways to interpret data- · Quantitative Data Interpretation · Qualitative Data Interpretation Qualitative Data Interpretation · Qualitative data tells us about the emotions and feelings of people. · Qualitative data interpretation is focused on insights and motivations of people. QUALITATIVE DATA COLLECTION METHODS Qualitative data tells us the stories behind the numbers, delving into people's experiences, opinions, and emotions. To gather this rich information, researchers use various methods: 1. Unearthing Existing Treasures: Record Keeping Think of record keeping as visiting a library. Researchers use existing reliable documents and data sources like: · Customer reviews: Analyzing online reviews helps understand customer satisfaction and areas for improvement. · Social media comments: Public social media comments can reveal public perception of a brand or product. · Surveys with open-ended questions: Open-ended questions in surveys allow participants to elaborate on their experiences. · Company documents: Internal reports, customer service logs, and meeting notes can provide valuable insights. 14 2. Direct Observation: Watching the Play Unfold Sometimes, the best way to understand someone is to observe them. Here's where direct observation comes in: · Ethnography: Researchers immerse themselves in a culture or setting to observe behavior firsthand. Imagine studying how families interact with technology in their homes. · Behavior observation: Observing participants in a controlled environment can reveal subconscious behaviors. For example, observing how customers navigate a store layout. 3. Deep Dives: Case Studies Case studies involve in-depth examinations of specific situations or individuals. It's like taking a magnifying glass to a single data point: · Customer case studies: Studying how a particular customer interacts with a product can reveal usage patterns and pain points. · Medical case studies: Examining individual medical cases can provide insights into rare diseases or treatment responses. 4. Group Discussions: Brainstorming Together Focus groups bring people together to discuss a specific topic. It's like a facilitated brainstorming session: · Focus groups: A moderator guides a group discussion, encouraging participants to share their experiences and opinions on a product, service, or idea. 5. Tracking Changes Over Time: Longitudinal Studies Longitudinal studies involve collecting data from the same source repeatedly over time. It's like watching a story unfold: · Customer satisfaction tracking: Regularly surveying customers helps understand how their perception of a product or service changes over time. · Employee engagement tracking: Tracking employee feedback over time can reveal trends in morale and satisfaction. 6. One-on-One Conversations: In-Depth Interviews Interviews allow for a deep exploration of a person's experiences and opinions. It's like having a focused conversation: · In-depth interviews: A skilled interviewer asks open-ended questions to understand a participant's perspective on a topic. Activity Trend Analysis Purpose: · This activity will engage youth with longitudinal studies – a study conducted over a considerable amount of time to identify trends and patterns. · The ability to identify trends and patterns in datasets allows us to make informed decisions about different tasks in our lives. 15 Activity Guidelines Let's do a small activity based on Identifying trends. · Visit the link by scanning the QR Code: https://trends.google.com/trends/?geo=IN (Google Trends) · Explore the website · Check what is trending in the year 2022 – Global Ø Make a list of trending sports (top 5) Ø Make a list of trending movies (top 5) · Check what is trending globally in the year 2022 List of trending movies (top 5) List of trending athletes (top 5) Understanding Quantitative Data Interpretation Numbers are everywhere! From daily temperatures to website traffic, quantitative data surrounds us. But how do we make sense of it all? This guide dives into the world of quantitative data interpretation, equipping you to transform numbers into insights. What is Quantitative Data Interpretation? Imagine a social media campaign. You want to know how many people liked your post. That's quantitative data – information expressed in numbers. Quantitative data interpretation is the process of analyzing this numerical data to answer questions like: · How many? (e.g., Number of website visits) · When? (e.g., Daily sales figures) · How often? (e.g., Customer purchase frequency) Unveiling the Data: Collection Methods So, how do we gather this numerical goldmine? Here are some key methods: · Interviews: Structured interviews with specific questions can yield valuable quantitative data (e.g., Average household income). · Polls: These quick surveys, often with single questions, provide a snapshot of public opinion in numbers (e.g., Percentage of people who prefer online shopping). · Observations: Recording data during a specific timeframe paints a quantitative picture (e.g., Number of cars passing through a tollbooth per hour). · Surveys: Reaching a large audience through surveys allows researchers to collect vast amounts of quantitative data (e.g., Customer satisfaction ratings on a scale of 1 to 5). · Longitudinal Studies: Conducted over extended periods, these studies track changes in numerical data over time (e.g., Monitoring a company's stock price over a year). 16 Let's summarize Qualitative and Quantitative data interpretation Qualitative & Quantitative Data Interpretation Qualitative Data Interpretation Quantitative Data Interpretation Categorical Numerical Provides insights into feelings and emotions Provides insights into quantity Answers how and why Answers when, how many or how often Methods – Interviews, Focus Groups Methods – Assessment, Tests, Polls, Surveys Example question – Why do students like Example question – How many students attending online classes? like attending online classes? Making Sense of the Numbers: Steps to Data Analysis Once you have your data, it's time to unlock its secrets! Here's a roadmap to guide you: 1. Identify the Variables: Are you dealing with categories (like favorite color) or numerical values (like exam scores)? Understanding the type of variables helps choose the right tools. 2. Descriptive Statistics: Summarize your data using tools like mean (average), median (middle value), or mode (most frequent value). 3. Measurement Scales: Data can be measured on different scales – ratio (has a true zero point, like height), interval (equal distances between values, like temperature), or ordinal (ranked categories, like customer satisfaction). Choosing the right scale ensures accurate interpretation. 4. Data Presentation: Now it's time to showcase your findings! Choose the most effective format: Types of Data Interpretation. There are three ways in which data can be presented o Textual DI: Brief written descriptions work well for small datasets (e.g., "70% of customers rated the product excellent"). o Tabular DI: Organize your data in a table with rows and columns for easy comparison (e.g., A table showing website traffic by month). o Graphical DI: Visualize your data using charts like: § Bar Graphs: Ideal for comparing categories (e.g., Comparing sales figures for different product categories). § Line Graphs: Reveal trends over time (e.g., A line graph tracking website traffic growth over a year). § Pie Charts: Show proportions of a whole (e.g., A pie chart depicting customer demographics by age group). 17 Examples in Action! Let's see quantitative data interpretation in action: Example 1: Exam Scores o Data: Scores of 50 students in a math test (ranging from 50 to 100). o Interpretation: We can calculate the average score (mean) to see how the class performed overall. The median score would tell us the middle value, and the mode would reveal the most frequent score. Example 2: Customer Satisfaction Survey o Data: Responses to a survey asking customers to rate their satisfaction on a scale of 1 (very dissatisfied) to 5 (very satisfied). o Interpretation: We can create a bar graph to compare the number of customers in each satisfaction category. Activity Visualize and Interpret Data Duration: 40 Minutes Purpose · This activity will engage youth with data visualization and interpretation · visualization makes it easier for us to extract useful information contained in the dataset Activity Guidelines · The table shows the details of a class consisting of 50 students and their scores ranging in the listed categories for 5 subjects: Math, Physics, Chemistry, Social Science, and Biology 18 Copy the table in an Excel sheet and create the following visualizations for the given data: · Make a bar graph showing the marks distribution for all 5 subjects · Make a pie chart showing the marks distribution for Physics · Make a line chart displaying the marks distribution for Chemistry Importance of Data Interpretation 19 CONCEPTUAL SKILLS ASSESSMENT A. Multiple Choice Questions (MCQs) 1. Data that is organized into rows and columns is classified as: a. Textual Data b. Structured Data c. Numeric Data d. Unstructured Data 2. Which of the following is NOT a method for data pre-processing? a. Error Correction b. Data Visualization c. Formatting d. Removing Duplicates 3. What are the basic building blocks of qualitative data? a. Individuals b. Units c. Categories d. Measurements 4. Which among these is not a type of data interpretation? a. Textual b. Tabular c. Graphical d. Raw data 5. A Bar Graph is an example of? a. Textual b. Tabular c. Graphical d. None of the above 6. _____________ relates to the manipulation of data to produce meaningful insights. a. Data Processing b. Data Interpretation c. Data Analysis d. Data Presentation B. Fill in the Blanks Transparency, Numeric, Data Acquisition, Continuous, Data Augmentation, Independent, Textual, Accurate, Data Feature, Deletion 1. Data that can be counted or measured is called __________________ data. 2. The process of gathering data is called __________________. 3. Data that consists of words and phrases is __________________ data. 4. Data that can take any value within a range is __________________ data. 5. Ensuring people understand how their data is used is an aspect of data __________________. 6. Removing entries with missing information is a data cleaning technique called __________________. 7. Creating new data points from existing data is called __________________. 8. Information about each piece of data within a dataset is called a __________________. 20 9. Features used to make predictions in an AI model are called __________________ features. 10. Data that is free from errors and reflects real-world values is considered __________________. C. State true or false 1. Textual data is always subjective and cannot be analyzed quantitatively. 2. Data acquisition only involves collecting new information. 3. Data cleaning always involves removing entries with missing information. 4. Data features are independent of each other and never influence each other 5. Ethical considerations for data acquisition are only relevant for personal data. 6. Data visualization is the final step in the data analysis process. 7. An AI model for facial recognition would primarily use numeric data 8. Data accuracy is not a concern for AI development, as powerful algorithms can compensate for errors. D. Short Answer Questions 1. Describe the difference between continuous and discrete data. 2. Explain two challenges of data acquisition for AI projects. 3. Why is it important to consider ethical principles when using data for AI development? 4. Briefly explain the concept of data augmentation. (2 marks) E. Competency-Based Questions 1. You are tasked with collecting data on customer satisfaction for a new e- commerce website. What methods would you use and why? 2. Imagine you're building an AI model to predict flight delays. What data features would you consider including? Explain your choices. 3. You're working on a project that uses social media data to understand public opinion on a new policy. How would you ensure responsible data collection practices? F. Reasoning & Assertion Directions: For each statement, decide whether the assertion (A) is the reason for the claim (R). 1. Assertion (A): Data cleaning is an essential step in data analysis. Reason (R): It ensures accurate and reliable results. (Correct) 2. Assertion (A): Data visualization is always the final step in the data analysis process. Reason (R): It helps communicate insights and identify patterns. (Incorrect - Visualization can occur throughout the process) 21
