Unit-2 (Research, Writing & Data)

Course Overview

This course connects three practical areas: AI for academic research, AI-assisted Excel and data understanding, and AI-powered data analysis. It moves from finding and writing about information to organizing, analyzing and interpreting data.

Module 4 — AI for Research & Academic Work

Module 5 — AI in Excel & Data Understanding

Module 6 — Data Analysis Using AI

Responsible and ethical AI use throughout the workflow

Learning Outcomes

Explain how AI supports academic research and literature review.

Use AI for writing assistance, summarization, citation support and idea generation.

Create specific academic prompts with audience, format, length, tone and source requirements.

Distinguish structured from unstructured data.

Understand Excel workbooks, worksheets, rows, columns and cells.

Recognize common Excel functions and AI-assisted features.

Apply the five stages of data analysis: collection, cleaning, processing, visualisation and interpretation.

Choose appropriate visualisation types for different analytical questions.

Identify AI benefits, limitations, hallucinations, bias, privacy and security risks.

Use AI as an assistant while preserving independent thinking and human verification.

MODULE 4 — AI FOR RESEARCH & ACADEMIC WORK

1. What Is Academic Research?

The presentation defines academic research as a systematic process of collecting information, investigating topics, analyzing evidence, discovering new knowledge and solving real-world problems.

Basic Research — focuses on foundational knowledge, such as studying climate patterns.

Applied Research — focuses on solving practical problems, such as renewable energy.

Quantitative Research — uses numerical data and statistical analysis.

Qualitative Research — uses observations, interviews and textual analysis.

2. Traditional vs AI-Powered Research

Traditional work often involves manual library searching, reading large volumes, organizing references and writing reports from scratch.

AI-powered research can support multi-database searching, paper summarization, citation generation, writing/editing, pattern detection and rapid literature reviews.

Key takeaway: AI is positioned as a research assistant that handles groundwork so researchers can spend more time on analysis, insight and creativity.

3. Information Gathering & Literature Review

AI can support searching across:

Research journals and papers

Academic databases

Books and online articles

Institutional repositories

Literature-review tasks supported by AI include finding relevant papers, summarizing key findings, identifying research gaps, and organizing/cross-referencing sources.

Key takeaway: The presentation frames the potential time shift as weeks of manual work becoming hours of AI-assisted groundwork.

4. AI Academic Writing Support

AI can improve:

Grammar and spelling

Vocabulary and word choice

Sentence structure

Readability and clarity

Formal or academic tone

AI can also help generate titles and outlines, abstracts and introductions, research summaries, draft essays/reports, presentation slides, assignments and answers.

5. The 7-Step AI Research Workflow

1. Topic Selection — AI suggests trending topics and possible gaps.

2. Information Collection — AI searches journals and databases.

3. Literature Review — AI summarizes existing studies.

4. Data Collection — Use surveys, experiments or observations.

5. Data Analysis — AI can detect patterns and trends.

6. Writing & Formatting — AI assists with drafting and formatting reports.

7. Citation & Plagiarism — AI can support reference checking and originality checks.

6. Citation & Referencing

APA — presented as commonly used in Social Sciences.

MLA — presented as common in Literature & Humanities.

Harvard — presented as used in international academia.

Chicago — presented as used in History & Publishing.

AI referencing benefits described in the presentation include speed, reduced formatting errors, centralized organization and flexibility between citation styles.

7. AI Tools for Research & Academic Work

ChatGPT — explanations, summaries, draft generation and brainstorming.

Grammarly — grammar correction, writing enhancement and tone improvement.

QuillBot — paraphrasing, summarization and sentence rewriting.

Google Scholar — scholarly articles, journal papers and academic citations.

Semantic Scholar — AI-powered discovery, recommendations and filtering.

Zotero / Mendeley — reference management, citation generation and paper organization.

8. Prompt Engineering for Research

Weak prompt: “Write about AI.”

Stronger prompt: “Write a 500-word academic explanation of AI in healthcare with real-world examples and formal tone.”

Be specific and detailed.

Define the output format.

Mention the audience level.

Add word or length constraints.

Specify formal/academic tone when appropriate.

Request citations when needed.

Key takeaway: A research prompt should reduce ambiguity rather than leave the AI to guess the intended scope.

9. AI for Summarization

AI can process research papers, academic journals, annual reports, theses and books into concise summaries.

Extractive summarization selects important sentences directly from the source.

Abstractive summarization creates a new, simplified explanation in different words.

Useful summary elements include key arguments, main findings, core conclusions and essential data.

10. AI for Data Analysis in Research

Organizing Data — structures raw datasets.

Pattern Detection — identifies hidden trends and correlations.

Visualizations — supports charts, graphs and dashboards.

Statistical Analysis — can assist with regression, classification and clustering.

11. Plagiarism Detection

The presentation describes plagiarism as copying others' work without proper credit. AI-based systems can compare documents against large databases, generate similarity scores and highlight matching passages.

Document is submitted.

Text patterns are tokenized/indexed.

Content is compared against online sources.

Similarity scores are generated.

Potential matching passages are highlighted.

Key takeaway: The presentation also notes false positives and missed paraphrased copying; human review remains necessary.

12. Translation & Multilingual Research

Read foreign research in another language.

Translate academic documents.

Collaborate across international research teams.

Support cross-border educational programs.

Improve access to global scientific discoveries.

13. Ethical Use of AI in Academics

Use AI to assist learning, not replace learning.

Verify facts, statistics and citations.

Avoid full dependency; critical thinking and independent analysis remain essential.

Give proper references to AI-assisted work and original sources.

Do not submit AI-generated work as your own when that violates academic integrity policies.

Protect privacy; do not enter sensitive personal or institutional data into public AI tools.

14. AI Hallucinations in Research

The presentation defines hallucination as incorrect or fabricated information presented with false confidence.

Fake citations — invented papers, authors or journals.

Wrong statistics — fabricated numbers or percentages.

False attributions — incorrect quotes or authorship.

Incorrect dates — wrong publication or event dates.

Cross-check references manually.

Verify sources in trusted databases.

Use Google Scholar to confirm papers.

Never blindly trust AI output.

Apply critical thinking.

15. Benefits, Limitations & Collaboration

Benefits: faster research, better organization, improved writing, increased productivity, better accessibility and 24/7 availability.

Limitations: inaccurate information, limited critical thinking, bias, privacy risks and overdependence.

Collaborative research: document sharing, workflow organization, multilingual translation and discussion summaries.

16. Future of AI in Research & Education

AI writing assistants are already mainstream for grammar, citations, summarization and drafting.

Autonomous research agents are presented as an emerging direction for searching, reading and synthesizing literature.

AI tutors may personalize curriculum, pace and difficulty.

Multimodal AI may combine text, images, audio and video.

Scientific-discovery systems may assist with hypotheses and experiments.

MODULE 5 — AI IN EXCEL & DATA UNDERSTANDING

17. Excel Fundamentals

Microsoft Excel is described as spreadsheet software for storing, organizing and analyzing data. AI enhancements can automate calculations, detect patterns and provide smart suggestions.

Workbook — the complete Excel file (.xlsx), containing one or more worksheets.

Worksheet — a single tab/page within a workbook.

Cell — the individual box at a row/column intersection; examples include A1 and B3.

Row — horizontal line of cells numbered 1, 2, 3, ...

Column — vertical line of cells labeled A, B, C, ...

18. Common Excel Functions

SUM(range) — adds numbers in a range. Example from the presentation: =SUM(A1:A10) → 550.

AVERAGE(range) — calculates the mean. Example: =AVERAGE(B1:B5) → 42.

MAX(range) — returns the largest value. Example: =MAX(C1:C20) → 98.

MIN(range) — returns the smallest value. Example: =MIN(D1:D10) → 3.

COUNT(range) — counts numerical entries. Example: =COUNT(E1:E50) → 48.

19. AI Features in Excel

Analyse Data suggestions — identifies patterns and proposes charts and summaries.

Automatic chart generation — recommends chart types based on data structure.

Formula recommendations — suggests formulas based on context.

Data cleaning — detects duplicates, inconsistencies and formatting errors and can offer fixes.

20. Understanding Data

Data may include numbers, text, images, audio, video and more. Understanding the data type is foundational to analysis.

Structured Data — organized in rows and columns; examples include Excel spreadsheets, SQL databases, financial records and attendance lists.

Unstructured Data — has no predefined organization; examples include emails, social media posts, videos, images and audio.

The presentation states that 80–90% of the world's data is unstructured and describes NLP and Computer Vision as ways of processing it.

MODULE 6 — DATA ANALYSIS USING AI

21. What Is Data Analysis?

Data analysis is the systematic process of examining datasets to draw conclusions about the information they contain. The presentation positions AI as a way to make this process faster, smarter and more accessible.

22. The 5 Steps in Data Analysis

1. Data Collection — Gather raw data from surveys, sensors, databases, web scraping or manual entry. AI can automate ingestion from multiple streams.

2. Data Cleaning — Remove errors, duplicates and inconsistencies. AI can detect missing values, outliers and formatting issues.

3. Data Processing — Transform raw data through aggregation, normalization, encoding and structuring.

4. Data Visualisation — Convert processed data into charts, graphs, heat maps and dashboards.

5. Interpretation — Draw conclusions from visualized data; AI can assist with anomaly detection, trend identification and natural-language summaries.

23. How AI Helps in Data Analysis

Detects trends and patterns across very large datasets.

Automates complex calculations.

Predicts future outcomes using historical trends and patterns.

Reduces repetitive human effort so analysts can focus on interpretation and strategy.

24. Choosing Data Visualisations

Bar Chart — compare quantities across categories, such as sales by product.

Pie Chart — show proportions or percentage distribution of a whole.

Line Graph — display trends and changes over time.

Histogram — show frequency distribution of numerical data within ranges/bins.

25. Real-World Applications

Business Forecasting — sales/revenue prediction, inventory and supply-chain optimization, market-trend detection and strategic planning.

Healthcare Analytics — medical-image diagnosis support, readmission-risk prediction, personalized treatment support and drug-interaction detection.

Marketing Analysis — customer behavior analysis, audience segmentation, campaign ROI measurement and product recommendations.

Financial Predictions — fraud detection, risk assessment, market forecasting and loan-approval optimization.

26. Automation in Data Analysis

Report Generation — automatically create detailed performance reports.

Trend Analysis — continuously monitor patterns without manual effort.

Dashboard Updates — refresh dashboards and synchronize data.

Data Classification — automatically sort and categorize incoming data.

Key takeaway: Automation can increase efficiency and reduce workload, but outputs still need appropriate human review.

27. Ethical Issues in AI Data Analysis

Privacy — sensitive personal data must be protected throughout processing.

Bias — biased training data can produce unfair or discriminatory outputs.

Transparency — AI decisions should be explainable and understandable to stakeholders.

Security — data systems need robust protection against cyber attacks and breaches.

28. Hallucinations & Incorrect Insights in Data Analysis

An AI hallucination is described as an incorrect conclusion, fabricated fact or false information that appears credible.

Verify outputs manually before acting.

Cross-check findings against multiple independent trusted sources.

Use reliable, verified datasets.

Continuously monitor and audit AI systems.

Key takeaway: Human oversight remains essential: AI is a tool, not a replacement for critical thinking.

29. Advantages of AI in Data Analysis

Faster processing of massive datasets.

Higher accuracy by reducing manual calculation and interpretation errors.

Predictive capabilities using historical patterns.

Scalability for very large datasets.

Automation of repetitive work.

30. Limitations of AI Data Analysis

Data dependency — AI needs high-quality, clean and sufficient data.

High costs — advanced infrastructure and maintenance can require significant investment.

Bias problems — poor or unrepresentative datasets can produce skewed results.

Lack of human judgment — AI cannot fully replicate human intuition, creativity or contextual reasoning.

Security risks — sensitive data may be vulnerable to breaches.

31. Future of AI in Data Analysis

More autonomous data analysis with fewer human prompts.

Instant, real-time insights on demand.

Prediction of complex behaviors across industries.

Multimodal analytics combining text, image, audio and video.

Conversational data analysis through natural language.

Emerging trends highlighted in the source: conversational analytics, autonomous BI systems and multimodal analytics.

FINAL PRACTICE & KNOWLEDGE CHECK

Scenario 1 — Academic Research

You are asked to investigate a research topic, summarize ten papers, identify gaps, draft a literature-review section and provide references. Design a responsible AI workflow using the course concepts.

Start with a specific topic and research question.

Use AI to locate and organize relevant literature.

Summarize papers while checking the original sources.

Identify gaps rather than treating AI suggestions as established facts.

Draft only after defining audience, tone, structure and length.

Verify every citation and important factual claim.

Scenario 2 — Excel & Data

You receive a spreadsheet containing attendance records with duplicate rows, missing values and inconsistent formatting. Describe the appropriate AI-assisted analysis sequence.

Inspect the workbook, worksheet, rows, columns and cells.

Clean duplicates, missing values and formatting inconsistencies.

Process the dataset into an analysis-ready structure.

Visualize attendance patterns using an appropriate chart.

Interpret trends and anomalies.

Verify AI-generated conclusions before using them for decisions.

Scenario 3 — Ethical Decision

An AI tool produces a convincing research statistic and a citation that cannot be found in the academic database. What should you do?

Do not use the statistic or citation as verified evidence.

Check the original source and trusted academic databases.

Search for the cited paper independently.

Replace or remove unsupported claims.

Use human judgment and document reliable sources.

Quick Revision Checklist

□ I can explain AI's role in academic research.

□ I can distinguish basic, applied, quantitative and qualitative research.

□ I can build a strong academic prompt.

□ I understand extractive vs abstractive summarization.

□ I can explain plagiarism detection and its limitations.

□ I can identify responsible AI practices in academics.

□ I know workbook, worksheet, row, column and cell.

□ I can explain common Excel functions.

□ I can distinguish structured and unstructured data.

□ I know the five data-analysis steps.

□ I can choose an appropriate chart for a data question.

□ I can explain AI automation, benefits, limitations and ethical risks.

Course Summary

Across research, Excel and data analysis, AI is presented as an accelerator for searching, organizing, writing, summarizing, calculating, visualizing and identifying patterns. The central learning principle is that AI improves speed and productivity but does not remove the need for verification, ethical practice, privacy protection, critical thinking and human judgment.