Visit the Dashboard: https://osotsia.github.io/svelte-d3-dashboard/
This is a self-contained, interactive web application for exploring the drivers of Levelized Cost of Heat. LCOH is the lifetime (projected) cost of a heat-generating solution divided by the total amount of heat it will generate. It is designed for techno-economic analysts to run scenarios, understand parameter sensitivities, and evaluate the performance of a surrogate machine learning model.
For ease of use, this entire application is compiled into a single, portable HTML file with no external dependencies, with the option of later integrating a python backend.
Features
Interactive Scenario Analysis
- LCOH Calculator: Sliders for real-time parameter adjustments and impact analysis.
- Scenario Management: Save, load, and compare named scenarios.
Sensitivity & Driver Analysis
- Sobol Indices: Quantify input impact on model variance (direct vs. total effects).
- Parallel Coordinates Plot: Visually filter and identify input/output correlations.
Surrogate Model Diagnostics
- Model Card: Key performance metrics (R², RMSE) and metadata.
- Permutation Importance: Rank parameter influence on surrogate model predictions.
- Residuals Plot: Diagnose model fit and systematic bias.
Data & Assumptions
- Document Assumptions: Log and display core model inputs, sources, and justifications.
- Track Benchmarks: View model outputs alongside external data for comparison.
Customizable Reporting
- Pinboard System: Pin any analysis, assumption, or benchmark to a central report.
- Analyst Notebook: Add qualitative summaries and notes to each scenario.
FAQ
Q1: What is the primary objective of the LCOH Workbench project? The objective is to accelerate techno-economic analysis and improve report consistency. It integrates a real-time calculator, scenario management, and visualizations into a single interface, consolidating workflows previously split across spreadsheets and model scripts.
Q2: Why was Svelte 5 chosen for the front-end framework?
Svelte 5 was chosen for code conciseness, performance, and a small bundle size, suitable for data-intensive applications. It compiles components to efficient JavaScript, and its rune-based reactivity ($state, $derived) simplifies state management, reducing code complexity and potential errors.
Q3: Can I upload my own dataset or connect a different model? No. All model logic and analytical data are pre-compiled into the application. User-provided data or models would require a future version with a backend service.
Q4: Where is my scenario data stored?
Scenario data is saved only in your browser's localStorage and is not sent to a server. This ensures privacy but means data is not accessible on other devices and will be lost if browser data is cleared.
Q5: How can I export a report from the Workbench? The application lacks a dedicated export function. Use your browser’s "Print" command and select "Save as PDF" to generate a static document of the Report view, which includes your pinned analyses and comments.
