Spatial evidence for better urban decisions.

The absolute view emphasizes differences in the overall scale of population and employment, while the relative view emphasizes differences in composition. Cells shown in the same color belong to the same cluster, suggesting that they have similar demographic, housing, and employment attributes. Around downtown, this separates job-intensive cells from the surrounding residential context instead of treating the entire area as one uniform center-city type.

Each 2 km grid cell is described using population, income, housing, resident-worker, workplace-job, and industry variables. The variables are standardized, UMAP reduces the multivariable profiles to a lower-dimensional representation, and HDBSCAN groups cells that have similar attribute patterns.

The colors summarize similarity in the selected data; they do not mean that two places are identical. UMAP preserves relationships among attribute profiles rather than geographic distance, and the exploratory clusters should be checked against the original variables and local knowledge.

The output is a numbered map and a ranked list of ten candidate areas. A higher rank means that a tract is closer to the selected tract across the measured variables; it does not mean that the two areas are identical. When the distance-aware option is used, the score also gives preference to physically closer locations. This helps distinguish nearby alternatives from places that look similar in the data but are farther away.

First, Census codes that represent missing values are removed. Highly skewed variables, such as population or job counts, are log transformed so that a few very large values do not dominate the comparison. The variables are then standardized and organized into demographic, housing, resident-workforce, and workplace-employment groups. A weighted distance calculation produces the similarity ranking, with an optional geographic-distance adjustment.

This is an initial screening tool, not a prediction that a site is available or suitable. A highly ranked tract can still differ in land use, current vacancies, wages, travel time, employer quality, infrastructure, or parcel-level feasibility. Those factors require a separate review before making a location decision.

The map shows where the thirty matches are located and whether they are concentrated near San Antonio or distributed across Texas. The ranked comparison identifies the closest statistical matches and shows which variables most strongly support the first-ranked result. County coverage and the added contribution of ranks eleven through thirty are reported so that users can judge whether the search found genuinely broader evidence or simply more matches from the same few places.

The same missing-value treatment, transformation, standardization, and feature-group weighting are applied to every eligible Texas tract. The tool calculates each tract's multivariable distance from the selected San Antonio tract, sorts the results, and reports the top thirty. Diagnostics show the similarity score, geographic distribution, county representation, and the demographic, housing, and employment signals behind the ranking.

A Texas tract can be statistically similar to the selected San Antonio tract while operating in a different policy, land-use, infrastructure, labor-market, or competitive environment. The Top30 list identifies places worth investigating; it does not establish that the locations are interchangeable or that any one of them is appropriate for a final investment or site decision.