1. Cost driver analysis
Multiple regression: which factors explain total cost, forecast accuracy on a hold-out sample, and a what-if estimator.
RegressionGoogle SheetsCost estimationLaura Camila Guayacan Ardila
Construction cost analytics
One question, three approaches: what drives construction cost, and which projects deviate from what we would expect? The same dataset of 6,123 projects is analysed with regression, SQL and an interactive dashboard, in the language a project manager uses to make decisions.
Key findings
- Equipment is a largely fixed cost (about $100k per project), so it weighs 31.7% of cost on projects under 2,000 sq ft versus 7.2% above 10,000 sq ft.
- Cost per sq ft outliers drop from 423 to 205 once equipment is removed: most of the "deviation" is a scale effect.
- Weather and risk indices do not explain the outliers.
Projects
2. Cost variance and outlier analysis
SQL on BigQuery: which projects and segments deviate from normal unit cost, and why.
SQLBigQueryVariance analysis3. Project cost dashboard
Interactive estimate with a range and a plain-language interpretation of the result (the AI layer interprets, it does not calculate).
HTMLDashboardCost reportingSkills shown
Cost estimationCost controlBudget forecastingVariance analysisRisk assessmentBenchmarkingMultiple regressionSQL (BigQuery)Data visualisationDashboards (HTML)Stakeholder reporting