Visual Analytics
2 credits · 16 chapters · 70 flashcards
Pre mid-sem
- What visual analytics isNot startedThe DIKW ladder, what analytics means, and the one comparison the professor keeps coming back to: data visualisation answers what, visual analytics answers why. · 40 min · 27 steps · Session 1Quiz: 6 questions, 70% to pass
- Types of data, and how to present themNot startedMetric versus non-metric, the four measurement scales, frequency distributions by hand, and which classical chart belongs to which kind of variable. · 45 min · 25 steps · Session 1Quiz: 6 questions, 70% to pass
- Choosing the right chartNot startedState the goal, look at the data type, then pick. The four goals, the chart for each, the specialised charts, and the charts the textbook tells you to avoid. · 45 min · 23 steps · Sessions 1-2Quiz: 6 questions, 70% to pass
- Preattentive attributes, Gestalt and data-inkNot startedWhy some charts are understood instantly and others are work: iconic memory, the four preattentive attributes, the four Gestalt principles, and decluttering. · 50 min · 25 steps · Sessions 1-2Quiz: 6 questions, 70% to pass
- Nodes, edges and the shape of a networkNot startedGraphs, edge lists and adjacency matrices; directed against undirected; small world, scale free and homophily; density by hand; and when a graph picture is the wrong answer. · 50 min · 27 steps · Session 2Quiz: 6 questions, 70% to pass
- Centrality and prestigeNot startedFive ways to define importance in a network, and how to compute four of them by hand: degree, closeness, betweenness, degree prestige and proximity prestige. · 1 h 5 min · 27 steps · Session 2Quiz: 6 questions, 70% to pass
- Community detection, motifs, Gephi and VOSviewerNot startedThe five network analysis techniques, what a community and a motif are, and the two tools: Gephi for a network you own, VOSviewer for a literature you have exported. · 55 min · 29 steps · Sessions 3-5Quiz: 6 questions, 70% to pass
- Matplotlib and SeabornNot startedThe plotting libraries, the six Matplotlib skeletons the class taught, and the parameter names the paper asks you to write out: width, color, bins, explode, autopct, cmap, marker. · 45 min · 26 steps · Sessions 6-7Quiz: 6 questions, 70% to pass
- Preparing data for visual analyticsNot startedETL, the three pre-processing families, the unification rules, noise and the three binning methods by hand, the four ways to fill a missing value, and normalisation against standardisation. · 55 min · 31 steps · Sessions 6-8 (deck posted 3 September)Quiz: 6 questions, 70% to pass
- Visual analytics with text dataNot startedWhy words are hard data, the five steps of NLP and the six stages of NLG, the pipeline from tokenization to chunking, stemming against lemmatization, sentiment analysis, bag of words, and the four limitations of a word cloud. · 1 h · 40 steps · Sessions 9-10Quiz: 6 questions, 70% to pass
Post mid-sem
- Power BI: interface, views, filters and the data modelNot startedDesktop against Service against Mobile, the three views and four panes, Get Data, primary and foreign keys as the model's backbone, cardinality and cross-filter direction, the four-level filter hierarchy, and the basic visuals. Reconstructed: check against your class slides. · 50 min · 24 steps · Sessions 11-12Quiz: 6 questions, 70% to pass
- Power Query: shaping data before it reaches the modelNot startedApplied Steps as a repeatable recipe, data types, the four removals, split and merge, fill down, unpivot, Group By, append against merge and the join kinds, custom and conditional columns, Close and Apply. Reconstructed: check against your class slides. · 50 min · 25 steps · Sessions 13-14Quiz: 6 questions, 70% to pass
- Maps, slicers and advanced filteringNot startedData Category so a place name can be plotted, Map against Filled Map against Shape Map, the ambiguity problem, the four slicer styles and slicer sync, basic against advanced filter conditions, Top N, and drill-down against drill-through. Reconstructed: check against your class slides. · 45 min · 21 steps · Sessions 13-14Quiz: 6 questions, 70% to pass
- DAX: calculated columns, measures and the profit margin ratioNot startedRow context against filter context, when to store a column and when to write a measure, SUM against SUMX, CALCULATE, DIVIDE, and the sales, profit and profit margin ratio the outline names, worked by hand. Reconstructed: check against your class slides. · 55 min · 24 steps · Sessions 15-16 (titled Introduction to DAX in Power BI on the Drive deck's session plan)Quiz: 6 questions, 70% to pass
- Choosing the right visual, editing and sortingNot startedThe chart-selection rules applied in Power BI, switching visual type without losing the field mapping, the formatting decisions that matter, conditional formatting, sort by column, tooltips, themes and accessibility. Reconstructed: check against your class slides. · 45 min · 20 steps · Session 17Quiz: 6 questions, 70% to pass
- Dashboards, storytelling and telling the truthNot startedWhat a data dashboard is and the three taxonomies that classify one, know your audience and your message, Aristotle's rhetorical triangle and Freytag's pyramid, and the deceptive designs Camm names. Reconstructed: check against your class slides. · 55 min · 27 steps · Sessions 18-20Quiz: 6 questions, 70% to pass
Assessment
| Component | Weight | Note |
|---|---|---|
| Mid-Sem Exam | 20% | Pen and paper, scenario based. |
| End-Sem Exam | 30% | Pen and paper, scenario based. |
| Class Participation | 15% | Judged on quality and frequency of contributions. Regular attendance is mandatory to receive any credit at all. |
| Quiz | 15% | No quiz paper or quiz pattern appears anywhere in the material. |
| Individual Assignment | 20% | PLO2a. Four assignments were set for Section A; only Assignment 3 (VOSviewer, 100 points, due 21 Aug 2026) has a brief in the export. Section B was set six, all 100 points, one per Excel practice set. |
What the exam looks like
The professor stated the format in a Classroom announcement on 10 September 2026: the examination is pen and paper and not Excel based. A scenario is given, and you identify and explain the appropriate visualisation aspects for that scenario. Section B's copy of the same announcement adds four words Section A's does not: as solved on excel in class.
So the paper is a "which visual, and why" paper. Expect a business situation, a described dataset (variable types, how many categories, whether it is time ordered), then: which chart, which colour scheme, which preattentive attribute, which Gestalt principle, and what is wrong with the obvious alternative. The in-class practice workbooks are the likeliest scenarios, re-asked in prose.
The numericals are the network measures, all by hand: density, degree, degree prestige, proximity prestige, degree centrality, closeness centrality, betweenness centrality and the geodesic distance matrix. The two Centrality Question PDFs are fully worked models of exactly what those questions look like, and the class Assignment sheet has three more networks of 8, 8 and 12 nodes.
Python, Gephi, VOSviewer and Power BI are examined in writing: name the function, name the parameter, explain what a setting does, interpret a map that is printed for you. Expect definition and comparison questions too, straight from the two-column tables in the notes: data visualisation versus visual analytics, histogram versus bar graph, degree prestige versus proximity prestige, undirected versus directed, stemming versus lemmatisation, normalisation versus standardisation, calculated column versus measure.
What the professor expects
Class participation carries 15 marks and the outline is explicit: contributions are judged on quality and frequency, with emphasis on insightful comments, relevant questions and collaborative interaction, and regular attendance is mandatory to receive credit for class participation.
Slide 3 of every deck sets the classroom rules: phones are strictly prohibited and stay inside bags; laptops are not permitted by default and only when the faculty says so; a device used without permission is confiscated and submitted to the Office; "no excuses will be entertained"; proxy attendance means serious disciplinary action; and nobody enters or leaves once the session has started.
Before the network sessions you were told to install the latest VOSviewer from the official site, open it to confirm it works, install the Java SE Runtime Environment if required, and create an Elsevier/Scopus account on your personal email ID.
The outline's only prerequisite is that "understanding of the concept of primary key and foreign key are required and appreciated". That is what the Power BI data model in the second half rests on.
Slide 2 of every deck is a copyright notice: the material is collected from various sources for classroom learning and cannot be made publicly available. Everything in this portal is therefore paraphrased, and no slide image or long slide quotation is reproduced.
Books
- Jeffrey D. Camm, James J. Cochran, Michael J. Fry and Jeffrey W. Ohlmann - Data Visualization: Exploring and Explaining with Data with MindTap (prescribed; the Drive copy is the First Edition, 2022)
- Sharada Sringeswara, Purvi Tiwari and U. Dinesh Kumar - Data Visualization: Storytelling Using Data (reference; not present in source/)