READOS / Technology
Reporting, History and Analytical Data
Data warehouses, history, hierarchies, measures, lineage and analysis that turn records into trustworthy reports.
Start reading →Contents
- 01Turning business data into decisions: customer, operations and quality data for small firms, used responsiblyHow small businesses can capture and use customer, sales, operations and quality data with CRM and ERP tools, start from questions, protect privacy and security, and avoid common traps.
- 02Reporting databases and data warehouses: separating day-to-day systems from analysisReports run on live systems slow work and disagree with each other. How reporting copies, data warehouses and dimensional models give consistent figures, and how to start small.
- 03Keeping history in business data: revisions, effective dates and audit trailsMany systems store only what is true now. How to keep the history businesses need: what price applied, which revision was built and who changed what, without drowning in records.
- 04Events, snapshots and workflow facts in reporting databasesHow to choose reporting rows for events, periodic observations and workflow milestones, reconcile their populations and avoid confusing activity with outstanding work.
- 05Preserving historical classifications in reporting dimensionsHow to report by the classifications that applied when an event happened, distinguish corrections from changes, and assign late records to the right dimension version.
- 06Managing reference codes as changing business dataDesign lookup values with stable meaning, ownership, effective dates and migration rules so that changing codes do not silently rewrite operational history.
- 07Reporting through ragged and shared-parent hierarchiesDesign reliable roll-ups when reporting hierarchies have uneven depth, shared membership or changing parents, with explicit attribution and reconciliation rules.
- 08When reporting measures cannot be addedAvoid misleading report totals by distinguishing additive measures, snapshots, ratios and distinct counts, and checking the population behind each result.
- 09Materialised views and reporting freshnessDesign stored reporting results with clear refresh rules, source watermarks and reconciliation checks, including the limits of incremental maintenance.
- 10Tracing report figures through data lineageHow to trace a reported number through source records, transformations and loading runs, investigate disagreements and assess the impact of changing a data definition.
- 11Federated queries and conflicting data meaningsAssess federated database queries by reconciling identifiers, definitions, freshness and failure behaviour before treating separate sources as one dataset.
- 12When aggregate reports reveal individual recordsExamine how totals, overlapping groups and repeated releases can expose individual values, and design reporting controls that account for combined information.
- 13Interpreting association rules in transaction dataUse support, confidence and lift carefully when exploring transaction patterns, with attention to basket definitions, absent items, changing populations and validation.
- 14Nearest-neighbour search and meaningful similarityDesign similarity searches by choosing useful features, distance rules and filters, then assessing exactness, recall and the meaning of returned neighbours.
These topic-based reading collections are organised alphabetically; chapters may be read in any order.
← All books