Source guide, coverage and dataset conventions
This week uses Lectures 7.1–7.7 and the supplied slide decks for 7.3 and 7.7.
| Source | Coverage in these notes |
|---|---|
| 7.1 Examples to introduce dictionary | Temporary seen flags, unique keys, word-length frequencies, city counts |
| 7.2 Birthday dictionaries | Day/month keys,29 observed dates, ID lists and representation choices |
| 7.3 Pseudocode for dictionaries; slides | {}, assignment, lookup, keys, iteration, averages, first-occurrence updates, manual versus built-in isKey |
| 7.4 Customer patterns Part1 | Row-count unit, category map, profiles, coarse bins, feature choices |
| 7.5 Customer patterns Part2 | Absolute/Manhattan distance, all15 pairs, triangular table, nontransitive closeness, alternatives |
| 7.6 Introduction to dictionaries | Food by shop/customer, dynamic keys, preserving bill IDs, testing food-shop evidence |
| 7.7 Real-time examples; slides | FoodD pseudocode, duplicate flags, birthday ID lists/dictionaries, POS lists and nearest-earlier-noun matching |
Data sources reused: all 30 Scores cards, 65 Paragraph Words cards, all 30 original Shopping Bills with155 visually checked item rows, and the ten verified historical train cards introduced in Week 6. Names, categories and IDs follow the original cards. The five broad shopping groups are explicit; raw category labels remain available in the data.
Source corrections are visible rather than silently copied: complete Food shop totals are 73, 15, 9 rather than the spoken 72, 10, 9; the spoken profile-distance discussion has occasional arithmetic/table slips, so all 15 distances are recomputed from the five original profiles. Advaith and Mohith follow the supplied-card spellings. The small-basket boundary is stated explicitly. The source’s pronoun heuristic is taught with its actual incorrect-language matches and sentinel.
Miniature illustrations labelled illustrative or excerpt do not replace full-dataset calculations. A dictionary table’s row order is a reading aid; no pseudocode key order is promised. Original student and word IDs are zero-based, bill IDs begin at 1, and train-number strings keep leading zeros. Course dictionary access/isKey are assumed constant-time; list scans, list storage, nested loops and output generation still require work.