MFM 216 Big Data Analytics Notes cover all four units of the GGSIPU MBA (Financial Management) Semester 4 elective in a 49-page printed English book. The notes include 7 solved numericals, 7 short Python programs, 11 numbered figures, 13 answered questions and 55 additional practice questions using banking and finance examples. Topics include Hadoop and HDFS, regression, data visualization, K-means, decision trees, Naive Bayes, Apriori and recommendation systems. They follow the MBA (FM) scheme labelled 2021-22 onwards.
Written by Prateek Talwar, Founder, Unnati Educations · Reviewed by Prateek Talwar · Checked against the MBA (Financial Management) Scheme and Syllabus, 2021-22 onwards · Updated 14 October 2026 · Editorial policy
Scheme: MBA (Financial Management), 2021-22 onwards | Applies to: batches admitted from 2021-22 onwards, subject to subsequent revisions | Last recorded syllabus verification: 13 October 2026
MFM 216 Big Data Analytics at a glance: GGSIPU MBA (FM) Sem 4 credits, units and exam pattern
MFM 216 Big Data Analytics introduces large-scale data processing, predictive analytics, machine-learning methods and recommendation systems through banking and financial applications. It is listed as a two-credit Semester 4 elective, with a separate one-credit practical paper, MFM 218 Big Data Analytics Lab.
| Code and title | MFM 216 · Big Data Analytics (printed as MFM - 216 in the source syllabus mapping) |
|---|---|
| Course and semester | MBA (Financial Management), Semester 4 |
| Category and credits | Elective; 2 credits; L-2 |
| Marks | 100: 25 internal (15 class test + 10 other assessment) and 75 external |
| Pass mark | Not established by the supplied syllabus details; consult GGSIPU Ordinance 11 and current examination notifications |
| Applies to | Batches covered by the MBA (FM) scheme labelled 2021-22 onwards |
| End-term pattern | 75 external marks; detailed question-paper pattern not confirmed in the supplied scheme |
| Units and hours | 4 units, 28 teaching hours: 7 + 7 + 7 + 7 |
| Paired lab | MFM 218 Big Data Analytics Lab: 1 credit, T/P-2, 40 internal + 60 external |
The supplied Semester 4 scheme lists MFM 216 among the elective choices, alongside MFM 202 Project Dissertation and MFM 204 Project Management in the semester structure. Because the elective-slot summary lists three credits while MFM 216 carries two and MFM 218 carries one separately, students should confirm with their institution whether the theory and practical papers are taken together for elective-credit planning.
Official reference: GGSIPU MBA (Financial Management) Scheme and Syllabus. The supplied editorial mapping identifies MFM 216 on PDF page 84, the elective list on page 8 and assessment details on page 3.
MFM 216 Big Data Analytics Notes unit-wise: 7 solved numericals, 7 Python programs and 13 answered questions across 4 units
The printed book is listed as 49 pages. Pages 1–3 contain the cover and index, while pages 4–48 contain 45 pages of unit content. Its examples connect data analytics with lending, investing, fraud detection and banking operations.
MFM 216 Unit 1: Introduction to Big Data
Unit 1 (pages 4 to 16) explains the five Vs of Big Data, the development from data warehouses to Hadoop, best practices, value validation, banking applications, Loshin's data-throttling test and return-on-investment approaches to quantifying data value.
It also introduces HDFS, MapReduce and four NoSQL database families, including how distributed systems support storage and processing of large datasets.
Unit content: Two worked illustrations, four numbered figures, four answered questions and fourteen practice questions.
Quick look: MapReduce counting UPI transactions by city (from page 14)
- Input: Millions of records containing city and transaction amount are distributed across four processing nodes.
- Map: Each node emits a pair such as (Delhi, 1) or (Mumbai, 1) for every relevant record.
- Shuffle and sort: Records with the same city key are grouped for reduction.
- Reduce: The values are summed to produce transaction counts, such as (Delhi, 4,20,000).
This example counts transactions by city. To calculate total transaction value instead, the mapper would emit the transaction amount rather than the number 1.
MFM 216 Unit 2: Predictive Analytics
Unit 2 (pages 17 to 27) covers ordinary least squares, six regression assumptions, simple and multiple linear regression, interpretation of coefficients, R², t-tests, F-tests and variance inflation factors.
It also introduces visualization methods, interactive data analysis, Python and R, with finance-based examples such as estimating stock beta.
Unit content: Two solved numericals, three short Python programs, two numbered figures, three answered questions and thirteen practice questions.
Quick look: estimating a stock's beta by simple regression (from page 19)
| Result | Value | Interpretation |
|---|---|---|
| Slope (b) | 1.45 | Estimated beta: a 1-percentage-point change in market return corresponds to a 1.45-percentage-point change in predicted stock return. |
| Intercept (a) | 2.30 | Predicted stock return when market return is zero, using the example's percentage-point scale. |
| R² | 0.978 | 97.8% of the observed variation in stock returns is explained by this fitted regression. |
| Predicted return at X = 12% | 19.70% | 2.30 + 1.45 × 12 = 19.70%. |
Illustrative stock-return regression results
- Result
- Slope (b)
- Value
- 1.45
- Interpretation
- Estimated beta: a 1-percentage-point change in market return corresponds to a 1.45-percentage-point change in predicted stock return.
- Result
- Intercept (a)
- Value
- 2.30
- Interpretation
- Predicted stock return when market return is zero, using the example's percentage-point scale.
- Result
- R²
- Value
- 0.978
- Interpretation
- 97.8% of the observed variation in stock returns is explained by this fitted regression.
- Result
- Predicted return at X = 12%
- Value
- 19.70%
- Interpretation
- 2.30 + 1.45 × 12 = 19.70%.
A beta greater than one indicates greater estimated sensitivity to market movements in this regression; it does not guarantee the stock will outperform the market.
MFM 216 Unit 3: Mining Big Data using Machine Learning
Unit 3 (pages 28 to 38) introduces the knowledge discovery process, CRISP-DM, distance measures and five clustering families.
The worked problems include K-means clustering of six credit-card customers, an ID3 decision tree for loan default and Naive Bayes classification of a borrower. The unit connects these methods with banking applications and model evaluation.
Unit content: Three solved numericals, two Python programs, three numbered figures, three answered questions and fourteen practice questions.
Quick look: evaluating a credit model on 1,000 test loans (from page 34)
| Metric | Calculation | Result |
|---|---|---|
| Accuracy | (80 + 860) ÷ 1,000 | 94% |
| Precision | 80 ÷ 120 | 66.7% |
| Recall | 80 ÷ 100 | 80% |
Classification metrics from the supplied confusion-matrix example
- Metric
- Accuracy
- Calculation
- (80 + 860) ÷ 1,000
- Result
- 94%
- Metric
- Precision
- Calculation
- 80 ÷ 120
- Result
- 66.7%
- Metric
- Recall
- Calculation
- 80 ÷ 100
- Result
- 80%
In this example, the positive class is assumed to represent default. The model correctly identifies 80 of 100 actual defaults, but 40 of the 120 predicted defaults are false alarms. This illustrates why accuracy alone can be misleading for imbalanced lending datasets.
MFM 216 Unit 4: Association and Recommendation System
Unit 4 (pages 39 to 48) covers five association-rule measures and the Apriori algorithm using a bank-product basket containing savings accounts, credit cards, insurance and mutual funds.
The unit then explains collaborative filtering, content-based recommendation, knowledge-based recommendation and hybrid approaches, with a mutual-fund rating example.
Unit content: Two solved numericals, two Python programs, two numbered figures, three answered questions and fourteen practice questions.
Quick look: association rules from the itemset {S, C, I} (from page 41)
| Rule | Confidence | Lift | Decision |
|---|---|---|---|
| {S, I} → {C} | 100% | 1.20 | Passes the confidence threshold; positive association. |
| {S, C} → {I} | 75% | 1.125 | Passes the confidence threshold; positive association. |
| {C, I} → {S} | 75% | 0.90 | Passes the confidence threshold, but lift below 1 indicates negative association relative to the baseline. |
| {S} → {C, I} | 60% | 0.90 | Fails the 70% confidence threshold. |
Illustrative association rules at a minimum confidence threshold of 70%
- Rule
- {S, I} → {C}
- Confidence
- 100%
- Lift
- 1.20
- Decision
- Passes the confidence threshold; positive association.
- Rule
- {S, C} → {I}
- Confidence
- 75%
- Lift
- 1.125
- Decision
- Passes the confidence threshold; positive association.
- Rule
- {C, I} → {S}
- Confidence
- 75%
- Lift
- 0.90
- Decision
- Passes the confidence threshold, but lift below 1 indicates negative association relative to the baseline.
- Rule
- {S} → {C, I}
- Confidence
- 60%
- Lift
- 0.90
- Decision
- Fails the 70% confidence threshold.
A rule can meet a minimum confidence threshold without having lift greater than one. Confidence and lift answer different questions and should both be interpreted when evaluating a recommendation rule.
Six visual data analysis techniques for multi-dimensional finance data: excerpt from page 23 of the MFM 216 notes
Excerpt from page 23, Unit 2: Predictive Analytics; adapted for the web
Figure 2.2: Classification of visual data analysis techniques
Big Data may contain hundreds of variables, making it difficult to represent every relationship in a conventional two-dimensional chart. The notes classify visual data analysis techniques into six families.
1. Standard two-dimensional and three-dimensional displays
Bar charts, line graphs, scatter plots, histograms, box plots and three-dimensional surface plots provide familiar ways to display data. Their main limitation is that they show only a small number of dimensions at once.
2. Geometrically transformed displays
These techniques transform multidimensional records into a geometric representation. Scatter-plot matrices display pairwise relationships between variables. Parallel coordinates represent variables as axes and individual observations as lines crossing those axes.
For example, a bank can compare customer income, age, balance and credit score using parallel coordinates. Principal component plots can represent patterns in a lower-dimensional space.
3. Icon-based displays
Each record becomes an icon whose features represent different variable values. In Chernoff faces, facial features such as the eyes, mouth and nose can encode financial ratios. Star plots and stick figures are other examples.
4. Dense pixel displays
Dense pixel techniques use small coloured cells to represent individual values, allowing a large amount of information to fit into a compact display. One application is comparing long price or return series across many securities.
5. Stacked and hierarchical displays
Treemaps represent hierarchical data using nested rectangles. A financial-market treemap might group companies by sector, size rectangles by market capitalization and colour them by daily return. Dendrograms and sunburst charts are other hierarchical displays.
6. Graph and network displays
Graph displays represent entities as nodes and relationships as edges. Banks can use network analysis to investigate fund flows, circular transactions, linked accounts and potential fraud rings.
Choosing the appropriate visualization
Line charts suit time series; scatter plots show relationships between two numerical variables; parallel coordinates and scatter-plot matrices help examine multidimensional data; treemaps show hierarchies; and network diagrams represent relationships among entities.
For text such as financial news or social-media posts, a word cloud may provide an initial view of word frequency, although more detailed text analytics are needed for reliable interpretation.
Do the MFM 216 notes cover the full GGSIPU syllabus?
The supplied editorial mapping marks the listed topics in Units I–IV as covered. One stated limitation remains: IBM InfoSphere BigInsights, named in course outcome CO5, is not covered. Therefore, the notes should not be described as covering every course outcome without qualification.
| Official unit | Topics | Mapped notes | Coverage |
|---|---|---|---|
| I: Introduction to Big Data | Concepts and characteristics, evolution, best practices, value validation, use cases, applications, quantification of value and high-performance storage | Sections 1–9, pages 4–15 | Listed topics covered according to source |
| II: Predictive Analytics | Simple and multiple regression, interpreting coefficients, visualizations, interaction techniques and introductory Python/R | Sections 1–10, pages 17–26 | Listed topics covered according to source |
| III: Mining Big Data using Machine Learning | Machine-learning-based data mining, K-means, classification, decision trees, evaluation, Naive Bayes and financial applications | Sections 1–8, pages 28–37 | Listed topics covered according to source |
| IV: Association and Recommendation System | Association rules, Apriori, candidate rules, applications and collaborative, content-based and knowledge-based recommendation | Sections 1–9, pages 39–47 | Listed topics covered according to source |
MFM 216 syllabus topics mapped to the printed notes
- Official unit
- I: Introduction to Big Data
- Topics
- Concepts and characteristics, evolution, best practices, value validation, use cases, applications, quantification of value and high-performance storage
- Mapped notes
- Sections 1–9, pages 4–15
- Coverage
- Listed topics covered according to source
- Official unit
- II: Predictive Analytics
- Topics
- Simple and multiple regression, interpreting coefficients, visualizations, interaction techniques and introductory Python/R
- Mapped notes
- Sections 1–10, pages 17–26
- Coverage
- Listed topics covered according to source
- Official unit
- III: Mining Big Data using Machine Learning
- Topics
- Machine-learning-based data mining, K-means, classification, decision trees, evaluation, Naive Bayes and financial applications
- Mapped notes
- Sections 1–8, pages 28–37
- Coverage
- Listed topics covered according to source
- Official unit
- IV: Association and Recommendation System
- Topics
- Association rules, Apriori, candidate rules, applications and collaborative, content-based and knowledge-based recommendation
- Mapped notes
- Sections 1–9, pages 39–47
- Coverage
- Listed topics covered according to source
The supplied mapping identifies the subject on page 84 of the GGSIPU MBA (Financial Management) Scheme and Syllabus.
This is a source-based editorial coverage assessment, not a new independent verification of the official PDF.
How to write MFM 216 answers: step-marked numericals for regression, ID3, Apriori and Naive Bayes
Six of the 23 long-answer practice questions in the supplied notes are numerical problems. Showing the method clearly is important for producing a complete answer.
- Regression: Show the formula, substitution, estimated coefficients and fitted equation. Interpret slope, intercept and R² in the context of the problem.
- ID3 decision tree: Calculate initial entropy, conditional entropy and information gain. Show the selected split, tree and resulting IF–THEN rules.
- Apriori: Show candidate generation and pruning in sequence, such as C₁ → L₁ → C₂ → L₂ → C₃. Finish with support, confidence and lift.
- Naive Bayes: Write the priors and conditional probabilities, calculate class scores, normalise them if probabilities are required, and identify the predicted class.
- Value quantification: Include a small ROI calculation when discussing the financial value of a Big Data project.
The 2.5-mark and 12.5-mark question labels used in the notes are editorial practice categories, not a confirmed official GGSIPU examination pattern.
Worked example: Naive Bayes borrower classification (page 35)
A young, self-employed applicant is assessed for default risk. The supplied example uses the following class priors:
P(Default = Yes) = 0.4
P(Default = No) = 0.6
Using the conditional probabilities supplied in the worked example:
- Default score: 0.4 × 0.75 × 1.00 = 0.300.
- No-default score: 0.6 × 0.333 × 0.333 ≈ 0.0665.
- Total score: 0.300 + 0.0665 ≈ 0.3665.
- Normalised default probability: 0.300 ÷ 0.3665 ≈ 81.9%.
- Prediction: Default = Yes, because it has the larger posterior score.
The source rounds the no-default score to 0.067 and reports approximately 81.8%. The small difference arises from rounding intermediate values. This example assumes the conditional independence used by Naive Bayes; the result is illustrative and not a calibrated real-world lending probability.
MFM 216 exam corner: 13 answered questions and 55 practice questions by unit
| Unit | Answered questions | Short-answer practice | Long-answer practice |
|---|---|---|---|
| Unit 1 | 4 | 8 | 6 |
| Unit 2 | 3 | 8 | 5 |
| Unit 3 | 3 | 8 | 6 |
| Unit 4 | 3 | 8 | 6 |
| Total | 13 | 32 | 23 |
MFM 216 answered and practice-question counts
- Unit
- Unit 1
- Answered questions
- 4
- Short-answer practice
- 8
- Long-answer practice
- 6
- Unit
- Unit 2
- Answered questions
- 3
- Short-answer practice
- 8
- Long-answer practice
- 5
- Unit
- Unit 3
- Answered questions
- 3
- Short-answer practice
- 8
- Long-answer practice
- 6
- Unit
- Unit 4
- Answered questions
- 3
- Short-answer practice
- 8
- Long-answer practice
- 6
- Unit
- Total
- Answered questions
- 13
- Short-answer practice
- 32
- Long-answer practice
- 23
Total: 13 answered questions + 55 additional practice questions = 68 questions. The practice set includes 32 short-answer and 23 long-answer questions. The book also contains seven solved numericals, seven short Python programs and eleven numbered figures.
Practice questions are supplied without model answers. The following four examples show the answer points students should prepare.
One question per unit, with answer pointers:
- Unit 1: Explain HDFS architecture. Pointer: Discuss the NameNode, DataNodes, metadata, block storage, replication and heartbeats. The notes use 128 MB blocks and three replicas as an illustrative configuration.
- Unit 2: Explain interaction techniques and Shneiderman's mantra. Pointer: State “overview first, zoom and filter, then details-on-demand”; explain filtering, zooming, linking and brushing.
- Unit 3: Explain entropy and information gain. Pointer: Use H(S) = −Σ pᵢ log₂ pᵢ; show entropy of zero for a pure set and one bit for an equally divided binary set, then explain how ID3 selects the split with highest information gain.
- Unit 4: What is the cold-start problem in recommendation systems? Pointer: Explain the lack of interaction history for new users or items, then compare content-based, knowledge-based and hybrid approaches.
The separate MFM 216 important-questions page is planned but not confirmed as published in the supplied source.
Who writes and checks the MFM 216 Big Data Analytics notes?
- Prateek Talwar, author: Founder, Unnati Educations; Associate Company Secretary (ACS), ICSI, membership no. 58704, since 15 May 2019; B.Com, University of Delhi (School of Open Learning), 2012. Author profile
Prateek Talwar is also identified as the subject reviewer. The supplied editorial record states that each MFM 216 unit was mapped against the MBA (Financial Management) scheme labelled 2021-22 onwards.
The source also reports that five Python programs reproducing hand-solved examples were re-run and matched their printed outputs, including beta 1.45, R² 0.978 and a mutual-fund rating of 4.5. This is a source-reported check rather than a fresh execution of the original programs for this page.
The notes focus on finance applications so that algorithms are connected to lending, investment or fraud-analysis decisions.
Last updated: 14 October 2026. Last recorded syllabus verification: 13 October 2026. Read the editorial policy.
MFM 216 and the MFM 218 Big Data Analytics lab (6 listed lab topics), plus other GGSIPU MBA (FM) Sem 4 papers
Paired lab: MFM 218 Big Data Analytics Lab, carrying one credit with T/P-2 and a 40-internal/60-external marks split. The separate MFM 218 notes page is pending because its notes PDF has not yet been supplied.
The six listed practical topics connect with the theory notes as follows:
- MapReduce: Unit 1, section 9.4, pages 13–14, introduces distributed counting and the MapReduce model.
- Word count: Unit 1's MapReduce explanation provides the underlying processing method.
- Directory and file operations: Unit 1, section 9.3, page 13, introduces HDFS. The supplied notes do not include the shell commands required for a complete file-operations practical.
- Linear regression using Python or R: Unit 2, sections 6 and 10, pages 21–25.
- K-means and classification: Unit 3 includes numerical and program-based examples on pages 31 and 36.
- Association rules: Unit 4 includes the Apriori example and program around page 42.
The six topics are an editorial grouping of the practical activities described in the source. Students should use the separate MFM 218 practical requirements for actual lab submissions.
Other Semester 4 papers connect with Big Data Analytics:
- MFM 214 FinTech: Alternative-data lending and financial technology.
- MFM 208 Risk Management and Insurance: Risk analytics, value at risk and fraud detection.
- MFM 210 Treasury and Wealth Management: Investment analytics and robo-advisory applications.
| Code | Paper | Relationship |
|---|---|---|
| MFM 202 | Project Dissertation | Research and applied project work |
| MFM 204 | Project Management | Project planning and delivery |
| MFM 206 | Business Valuation | Financial analysis and valuation |
| MFM 212 | Sustainable Finance | Financial and sustainability applications |
Other papers identified in the supplied MBA (FM) Semester 4 scheme
- Code
- MFM 202
- Paper
- Project Dissertation
- Relationship
- Research and applied project work
- Code
- MFM 204
- Paper
- Project Management
- Relationship
- Project planning and delivery
- Code
- MFM 206
- Paper
- Business Valuation
- Relationship
- Financial analysis and valuation
- Code
- MFM 212
- Paper
- Sustainable Finance
- Relationship
- Financial and sustainability applications
MFM 216 Big Data Analytics notes FAQs
How do I get the full MFM 216 notes?
Use the WhatsApp order button below to request the full 49-page printed English book. The complete book is not sold as a downloadable PDF. A seven-page sample PDF is planned, but its live availability still needs confirmation.
Do I need to know coding to use the MFM 216 notes?
No advanced programming background is assumed in the supplied notes. The syllabus mapping introduces Python or R, and the book includes seven short Python programs of fewer than fifteen lines each, with comments explaining key steps. Five programs reproduce examples already worked by hand.
How is MFM 216 marked, and what is the passing mark?
The supplied scheme gives 100 marks: 25 internal and 75 external. The internal component is described as a 15-mark class test plus 10 marks for other assessment activities. A specific pass percentage is not established by the supplied details. Consult GGSIPU Ordinance 11 and current examination notifications for the applicable requirements.
Is MFM 216 a two-credit or three-credit elective?
MFM 216 Big Data Analytics is listed as two credits, with two lecture hours per week. MFM 218 Big Data Analytics Lab is listed separately as one credit. Because the Semester 4 elective-slot table shows three credits, confirm with your institution whether the two papers must be selected together.
Which MFM 216 unit has the most solved numericals?
Unit 3, Mining Big Data using Machine Learning, has three of the seven solved numericals: K-means clustering, an ID3 decision tree and a Naive Bayes classification. Units 2 and 4 have two solved numericals each. Unit 1 contains two worked illustrations instead.
My MBA (FM) batch joined in 2025-26. Do these notes match my syllabus?
The notes follow the MBA (Financial Management) scheme labelled 2021-22 onwards. A batch admitted in 2025-26 falls within that stated period unless a revised scheme applies. Compare the four unit headings with your institution's current syllabus before ordering.
Get the full MFM 216 Big Data Analytics Notes on WhatsApp
The full printed English book is listed as 49 pages, covering four units with seven solved numericals, seven Python programs, eleven numbered figures, thirteen answered questions and fifty-five additional practice questions. The source also reports four case studies.
| Part | Pages | Solved numericals | Python programs |
|---|---|---|---|
| Cover and index | 1–3 (3) | — | — |
| Unit 1: Introduction to Big Data | 4–16 (13) | 0 (2 illustrations) | 0 |
| Unit 2: Predictive Analytics | 17–27 (11) | 2 | 3 |
| Unit 3: Mining Big Data using Machine Learning | 28–38 (11) | 3 | 2 |
| Unit 4: Association and Recommendation System | 39–48 (10) | 2 | 2 |
| Final page to confirm | 49 (1; content not identified) | — | — |
| Listed book total | 49 | 7 | 7 |
MFM 216 printed notes: page allocation and worked content
- Part
- Cover and index
- Pages
- 1–3 (3)
- Solved numericals
- —
- Python programs
- —
- Pages
- 4–16 (13)
- Solved numericals
- 0 (2 illustrations)
- Python programs
- 0
- Pages
- 17–27 (11)
- Solved numericals
- 2
- Python programs
- 3
- Pages
- 28–38 (11)
- Solved numericals
- 3
- Python programs
- 2
- Pages
- 39–48 (10)
- Solved numericals
- 2
- Python programs
- 2
- Part
- Final page to confirm
- Pages
- 49 (1; content not identified)
- Solved numericals
- —
- Python programs
- —
- Part
- Listed book total
- Pages
- 49
- Solved numericals
- 7
- Python programs
- 7
The mapped content accounts for 48 pages, including the cover and index. Page 49 has not been assigned to a specific section in the supplied source. The complete notes are sold as a printed book, not as a PDF download.
Delivery: within 5 to 7 days at most; the exact time may vary with your location.
Order the MFM 216 notes book on WhatsApp (4 units, 49 pages)
Unnati Educations is an independent education provider based in Vikaspuri, Delhi. We are not affiliated with, authorised by, or endorsed by Guru Gobind Singh Indraprastha University (GGSIPU). 'GGSIPU' and 'IPU' are used only to describe the syllabus our notes follow. For official information, visit ipu.ac.in.
Ready for the full notes?
Discuss coverage and delivery with our team.