Numerical Discrepancy Investigation & Fixes | kawakamisekkei

When Excel, databases, old and new systems, or reports disagree on the same metric, we pinpoint where the discrepancy arises — JOIN conditions, duplicates, NULLs, master-data mismatches, definition differences — and fix it within an agreed scope, with before/after verification. A service by kawakamisekkei Inc. (Tokyo, Japan). Details and contact: https://kawakamisekkei.com/

Transcript

1. SERVICE 02 Numerical Discrepancy Investigation & Fixes Identifying why numbers differ across Excel, SQL, and reports — and fixing the cause kawakamisekkei Inc. — Tokyo, Japan
2. SERVICE 02 Do Any of These Sound Familiar? Excel results don't match the database Old and new systems show different numbers for the same metric Unexpected discrepancies appeared after changing SQL or aggregation logic After data migration, record counts, totals, or subject counts differ Duplicates, missing values, NULLs, or master-data mismatches are suspected Same subject, same metric — yet Excel and the database, old and new systems, or different reports disagree. We pinpoint where the discrepancy arises from existing documentation, SQL, data, and outputs. When the cause lies within the target SQL, Excel, or aggregation conditions, we also fix it and re-verify. © kawakamisekkei Inc. 02
3. SERVICE 02 What We Do PHASE 1 PHASE 2 Investigation Fix & re-verify Check aggregation targets, periods, and reference dates; units, Fix the cause when it lies within the target SQL, Excel, or aggregation numerators / denominators, and rounding; JOIN conditions and the conditions. Compare results before and after, and document the handling of duplicates, missing values, and NULLs; master data, code verification method and its results. systems, and key fields; processing conditions, extraction timing, and post-output edits. Map the definition differences between the compared processes. © kawakamisekkei Inc. 03
4. SERVICE 02 Deliverables 01 Where the discrepancy occurs, the figures involved, the cause, and its impact 02 Organized aggregation definitions and processing conditions 03 SQL, Excel, or programs fixed within the agreed scope 04 Before / after comparison, with verification method and results © kawakamisekkei Inc. 04
5. SERVICE 02 Scope & Terms One report, one metric, or one aggregation process is the unit of engagement. Scope and verification method are fixed before the investigation begins. Redesigning the source database, large changes spanning multiple systems, and rebuilding the aggregation pipeline as a whole are out of scope and organized as separate work. What to share when you contact us Tell us the symptom — which numbers disagree — and how those numbers are produced. Please prepare the affected reports or outputs, together with the Excel sheets, SQL, and other logic that computes them. Data handling: Received data, SQL, Excel files, report formats, and aggregation definitions are used only for this engagement, never reused elsewhere. Return or deletion after completion is agreed in advance. © kawakamisekkei Inc. 05
6. PROCESS How We Work 1 2 3 4 5 Discovery Baseline Delivery Acceptance Release / Operate Organize requirements, Fix what will be Implement only what was Feedback is handled Organize publication, constraints, and decision criteria; define the scope implemented, how it is verified, and how changes are handled fixed in the Baseline separately as fixes, additions, or premise changes update, and operating conditions for continued use Implementation scope is fixed before Delivery begins, and only confirmed items are implemented. Additions and premise changes are handled as the next round of scoping. Small, narrowly scoped engagements are welcome. © kawakamisekkei Inc. 06
7. COMPANY Company Profile We implement statistical models, Company business logic, and data infrastructure Representative in the Data Science Tech domain. Location Grounded in applied mathematics and software implementation, we convert the Founded recurring processes and decisions inside a business into reproducible code. Domain Website © kawakamisekkei Inc. kawakamisekkei Inc. （株式会社川上設計） Hiroshi Kawakami, Representative Director TM Hamacho Bldg. 6F, 1-9-1 Nihonbashi-Hamacho, Chuo-ku, Tokyo, Japan September 1, 1989 Data Science Tech https://kawakamisekkei.com/ 07
8. Contact Please reach us via the Contact form on our website. We start by reviewing your current workflow and materials, then propose a scope. https://kawakamisekkei.com/ © kawakamisekkei Inc.