This tutorial provides a fully self-contained implementation and explanation of the DL2 composite indicator algorithm. It introduces the theoretical foundation of DL2 as a weighted Minkowski norm, outlines its connection with modern composite indicator methodology, and demonstrates how the method integrates polarity-based normalization, powered distances, Fréchet-type aggregation, and machine-learning-based weighting via MARS. The document guides the reader step-by-step through the improved DL2 workflow, including convex weight enforcement, adaptive smoothing, and iterative convergence assessment. A simulated dataset is used to illustrate each component of the algorithm, offering a practical and reproducible example suitable for methodological learning, benchmarking, and applied research.

Transparency is a vital aspect of good governance that ensures accountability, integrity, and citizen participation. However, its measurement remains a challenge due to the ambiguous nature of human language. Traditional approaches rely on binary or categorical dimensions to measure transparency, which can lead to limited insights and inaccurate policy decisions. To overcome this challenge, we develop a transparency indicator based on fuzzy metrics. The transparency notion is contrary to the binary principle of having a property or not having it, that is, to be transparent or not. In our case, we apply a novel methodology designed for formative measurement and partially compensatory models using a set of qualitative dimensions. This tool provide a measure of how close (observations close to 1) or far (observations close to 0) each observation (in this setting municipality), is from the analyzed concept. This is calculated for each indicator preserving the metric structure. In this setting, this approach captures the complexity of transparency and can better identify areas for improvement in local governments. In a second step, we analyze the political, economic, and social determinants of this transparency index, relying on spatial econometric techniques to account for spatial dependencies and ensure the accuracy of our findings. In this work, we use a novel dataset that comprehensively measures transparency at the municipal level in the Valencian region. As far as we know, this is the first database that covers all municipalities in the region. By capturing the complexity of transparency, our approach can better identify areas for improvement and lead to more effective policy interventions. We argue that traditional approaches that rely on binary or categorical variables may oversimplify the concept, resulting in inadequate policy solutions that fail to address the real challenges faced by local governments.