Technical appendix

Engineering behind rodado

The public-facing site is a sociological reading of Brazilian public data. This page is the engineering appendix for readers interested in how the platform behind it is built — ingestion, ontology, AI query layer, and a direct comparison to how the same architecture maps onto Palantir Foundry.

Engineering competencies demonstrated

Each capability below is backed by a concrete, deployed piece of this project — not a claim, a running system.

Delivered case studies

Five shipped applications over Brazilian public data — each with a stakeholder problem, a data model, an application built, and a measurable outcome.

User workflows

Three analyst scenarios showing the full data → insight → decision arc.

Compliance — company integrity check

A compliance team verifies whether companies awarded public contracts have directors appearing in other sanctioned or flagged entities.

SELECT e.razao_social, e.cnpj, s.nome_socio,
       COUNT(DISTINCT e2.cnpj) AS other_entities,
       SUM(c.valor_contrato)   AS total_contracts_brl
FROM br_me_cnpj.estabelecimentos e
JOIN br_me_cnpj.socios s     ON e.cnpj_basico = s.cnpj_basico
JOIN br_me_cnpj.socios s2    ON s.cnpj_cpf_socio = s2.cnpj_cpf_socio
                             AND s2.cnpj_basico <> s.cnpj_basico
JOIN br_me_cnpj.estabelecimentos e2 ON s2.cnpj_basico = e2.cnpj_basico
JOIN br_cgu_compras_governamentais.contratos c ON e.cnpj = c.cnpj_contratado
WHERE e.sigla_uf = 'SP' AND c.ano = 2023
GROUP BY 1,2,3 HAVING other_entities > 2
ORDER BY total_contracts_brl DESC

Decision: flag companies for manual review; route to procurement governance team.

Policy — infrastructure gap prioritization

A state health secretariat identifies municipalities with critically low hospital bed coverage to prioritize federal budget allocation.

SELECT m.nome, m.sigla_uf, pop.populacao,
       ROUND(cnes.leitos_sus * 1000.0 / NULLIF(pop.populacao,0), 2) AS leitos_por_mil,
       ideb.nota_media  AS ideb_fundamental
FROM br_bd_diretorios_brasil.municipio m
JOIN br_ibge_populacao.municipio pop  ON m.id_municipio = pop.id_municipio  AND pop.ano  = 2022
JOIN ( SELECT id_municipio, SUM(leitos) AS leitos_sus
       FROM br_ms_cnes.estabelecimento WHERE ano = 2023
       GROUP BY id_municipio ) cnes   ON m.id_municipio = cnes.id_municipio
JOIN br_inep_ideb.municipio ideb      ON m.id_municipio = ideb.id_municipio AND ideb.ano = 2021
WHERE pop.populacao > 10000
ORDER BY leitos_por_mil ASC, ideb_fundamental ASC
LIMIT 50

Decision: top 50 ranked shortlist delivered to budget committee; top 10 flagged for emergency transfer.

Investigation — electoral spending anomalies

An investigative journalist tracks whether campaign spending correlates with post-election public contract awards for companies linked to the same candidate.

SELECT cand.nome_candidato, cand.sigla_partido,
       SUM(desp.valor_despesa)          AS campaign_spend,
       SUM(cont.valor_contrato)         AS post_election_contracts,
       COUNT(DISTINCT cont.cnpj_contratado) AS linked_companies
FROM br_tse_eleicoes.candidatos cand
JOIN br_tse_eleicoes.despesas_candidato desp ON cand.id_candidato = desp.id_candidato
JOIN br_me_cnpj.socios s     ON cand.cpf_candidato = s.cnpj_cpf_socio
JOIN br_cgu_compras_governamentais.contratos cont
     ON s.cnpj_basico = SUBSTR(cont.cnpj_contratado,1,8) AND cont.ano > cand.ano
WHERE cand.ano = 2022 AND cand.sigla_uf = 'SP'
GROUP BY 1,2 HAVING post_election_contracts > 1000000
ORDER BY post_election_contracts DESC

Decision: shortlist of 12 candidates forwarded to editorial team with source data for verification.

Domain ontology

Brazilian public data modeled as typed entities with explicit join keys.

Architecture

Data → semantic model → access control → application → users. Each layer independently deployable.

AI-powered query layer

Semantic table selection over 533 datasets prevents hallucination on a large schema.

$ ask "Qual o município com maior mortalidade infantil no Nordeste em 2021?"

→ SELECT m.nome, m.sigla_uf, s.taxa_mortalidade_infantil
  FROM br_ms_sim.municipio s
  JOIN br_bd_diretorios_brasil.municipio m ON s.id_municipio = m.id_municipio
  WHERE m.sigla_uf IN ('BA','PE','CE','MA','PI','RN','PB','AL','SE')
    AND s.ano = 2021
  ORDER BY s.taxa_mortalidade_infantil DESC
  LIMIT 1

Palantir Foundry mapping

This is not a Foundry deployment — it's an open-source system that reproduces the same architectural layers and mental model, for readers who know Foundry and want a quick translation.

Data quality & governance

Stack

Full engineering reference (environment variables, deploy, services, pipeline commands): see TECHNICAL.md in the repository.