Skip to content

Blog Posts

The spreadsheet that became a system

By Marco A. Parra on Aug. 21, 2026 · 6 min read

Between 2017 and 2019 an Excel workbook with macros became the specification of a dimensional-control system for small sawmills: Bluetooth caliper, mobile app and statistical analysis in the cloud. This article covers the method — four control points, nominal versus target, deviation charts — and the design decision that determines whether such a tool ever reaches a small company. No code, no plant data.

data-acquisitionoperationsspcwood-industry

Read More →

What 68 sawmills told us about their data

By Marco A. Parra on Aug. 21, 2026 · 7 min read

In 2019 I was first author of INFOR Technical Report No. 223: a survey of 68 small and mid-sized Chilean sawmills about their machinery and, above all, how they capture and use their data. This article walks through what we found — 79 % tracked lumber dimensions on paper forms — and why that diagnosis ended up as a software specification.

data-acquisitionindustry-4-0operationswood-industry

Read More →

Bridging a COBOL Production System Without Touching It

By Marco A. Parra on Aug. 20, 2026 · 5 min read

A COBOL plant system exported fragmented flat files every night and nobody could report on them. This post walks through the integration platform that turned those exports into a normalized relational model — idempotent ingestion, transactional boundaries, and a scheduling module built on top — and what it taught me about modernizing without a rewrite.

backenddjangoetllegacy-systemspostgresql

Read More →

Hybrid Metaheuristics for Job-Shop Tardiness: What Worked and What Did Not

By Marco A. Parra on Aug. 20, 2026 · 4 min read

Minimizing total tardiness in a job shop is NP-hard, and real plants add precedence constraints that classical benchmarks ignore. This post explains how GRASP, VNS and ILS were combined into a hybrid engine, how the results were validated statistically, and — just as important — what the study does not claim.

metaheuristicsoperations-researchoptimizationpythonscheduling

Read More →

Small Objects, Big Images: Evaluating YOLO and SAHI on 4K Drone Imagery

By Marco A. Parra on Aug. 11, 2026 · 4 min read

Standard YOLO inference silently loses small objects when high-resolution drone images are downscaled to the model's input size. This post walks through how the AgriDrone Vision pipeline uses SAHI sliced inference and a COCO-based evaluation layer to make that trade-off measurable — and why the real deliverable is a georeferenced map, not a bounding box.

computer-visionmachine-learningprecision-agricultureyolo

Read More →

Data Management and Visualization

By Marco A. Parra on Aug. 29, 2023 · 3 min read

This is a publication that provides data management for data exploration in the Mars craters dataset. Selected dataset: marscrate_pds This data set provides information about the morphological features of the surface of Mars. It contains data about recorded latitude, longitude, circular diameter, depth of rim floor, other morphological information, and the number of layers of craters on Mars.

pandas

Read More →