The tool

A PhD framework, turned into a web app

FRANCISCO (FRActal Network Cloud Infrastructure Service Comparison and Optimisation) is an R/Shiny application, open source under the Apache 2.0 license, that automates the entire analysis: from importing raw time series to the final ranking of virtual networks — without requiring the user to be an expert in mathematics, statistics or DEA.

Open the app ↗ Read the paper (CSBC 2023)

Usage flow

Four tabs, one click at a time

Home tab — upload the time series, choose decision variables and the fractal-dimension method, generate the DMU table.

DMU Analysis tab — exploratory analysis of each series: autocorrelation function (ACF), histogram, periodogram and log-log cumulative distribution (LLCD).

DEA Results tab — efficiency ranking. In super-efficiency models (SMDEA), the most efficient DMU appears first with a score above 1.

DEA Plots tab — efficiency frontier and unbounded-ray chart, generated automatically from the chosen DEA model.

Schematic illustrations of the flow described in Andrade & Marques Júnior (CSBC 2023) — not screenshots of the original application.

Under the hood

Technical stack

R + Shiny

Front-end and back-end in a single language, with HTML/CSS components via Shiny and shinyjs for JavaScript interactions.

Benchmarking

R package used to compute the classical DEA models and generate the efficiency frontiers.

fractaldim, pracma

Fractal dimension calculation (madogram, variogram, rodogram, box-count) and the Hurst parameter (R/S).

ptsuite, longmemo, TSA

Whittle estimator, heavy-tail alpha parameter, and time-series analysis.