How can the LAPD identify unique crime ‘fingerprints’ by integrating spatial hotspots, temporal patterns, and victim demographics to transition from reactive to proactive policing? This analysis aims to move from broad patrolling to Precision Policing. By segmenting crimes into specific “Risk Fingerprints,” we provide the LAPD with actionable intelligence to deploy specialized units where they are most needed.

# 220_3_PNAD_Income_per_decile_INDIVIDUAL_TOGGLE_AGGREGATES #============================================================================== # MELHORIAS EM RELAÇÃO AO 190_2: # 1) Cada decil é independente na legenda (sem legendgroup compartilhado) # 2) Grupos agregados: Bottom 40%, Bottom 50%, Bottom 80% # → Renda média ponderada do grupo inteiro por ano # 3) Grupos agregados com linha tracejada para distinguir dos decis #============================================================================== library(dplyr) library(survey) library(Hmisc) library(plotly) # --- CONFIGURAÇÕES DE VISUALIZAÇÃO --- COR_MEDIA_NACIONAL <- "#2ca02c" LARGURA_MEDIA_NACIONAL <- 4 TAMANHO_MARCADOR_MEDIA <- 12 MOSTRAR_IC_MEDIA_NACIONAL <- TRUE # Grupos agregados: nome, decis incluídos, cor GRUPOS_AGREGADOS <- list( list(nome = "<b>Bottom 40%</b> (D1–D4)", decis = 1:4, cor = "#e377c2", dash = "dash"), list(nome = "<b>Bottom 50%</b> (D1–D5)", decis = 1:5, cor = "#ff7f0e", dash = "dashdot"), list(nome = "<b>Bottom 80%</b> (D1–D8)", decis = 1:8, cor = "#9467bd", dash = "dot") ) # ============================================================================= # 1) PREPARAÇÃO DOS DECIS NACIONAIS # ============================================================================= pop_decis <- Salata_etal_2025_dataset %>% group_by(ano) %>% mutate( decil_renda = cut( renda_dom_pcta, breaks = Hmisc::wtd.quantile(renda_dom_pcta, weights = peso, probs = seq(0, 1, 0.1)), labels = paste0("D", 1:10), include_lowest = TRUE ) ) %>% ungroup() %>% filter(!is.na(decil_renda)) %>% mutate(decil_num = as.numeric(gsub("D", "", decil_renda))) anos_disponiveis <- sort(unique(pop_decis$ano)) # ============================================================================= # 2) CÁLCULO: DECIS + NACIONAL + GRUPOS AGREGADOS # ============================================================================= lista_decis <- list() lista_nacional <- list() # Uma lista por grupo agregado lista_agr <- lapply(seq_along(GRUPOS_AGREGADOS), function(x) list()) for (a in anos_disponiveis) { df_ano <- pop_decis %>% filter(ano == a) design_pop <- svydesign(ids = ~1, weights = ~peso, data = df_ano) # --- Por decil --- renda_decil <- svyby(~renda_dom_pcta, ~decil_renda, design_pop, svymean, vartype = "ci") renda_decil <- renda_decil %>% mutate(ano = a, decil = as.character(decil_renda), media = renda_dom_pcta, lower = ci_l, upper = ci_u) lista_decis[[as.character(a)]] <- renda_decil # --- Nacional --- m_nac <- svymean(~renda_dom_pcta, design_pop) ci_nac <- confint(m_nac) lista_nacional[[as.character(a)]] <- data.frame( ano = a, media = as.numeric(m_nac["renda_dom_pcta"]), lower = ci_nac["renda_dom_pcta", "2.5 %"], upper = ci_nac["renda_dom_pcta", "97.5 %"] ) # --- Grupos agregados --- for (g in seq_along(GRUPOS_AGREGADOS)) { df_grupo <- df_ano %>% filter(decil_num %in% GRUPOS_AGREGADOS[[g]]$decis) design_grupo <- svydesign(ids = ~1, weights = ~peso, data = df_grupo) m_g <- svymean(~renda_dom_pcta, design_grupo) ci_g <- confint(m_g) lista_agr[[g]][[as.character(a)]] <- data.frame( ano = a, media = as.numeric(m_g["renda_dom_pcta"]), lower = ci_g["renda_dom_pcta", "2.5 %"], upper = ci_g["renda_dom_pcta", "97.5 %"] ) } } dados_decis <- bind_rows(lista_decis) %>% mutate(decil_num = as.numeric(gsub("D", "", decil))) dados_nacional <- bind_rows(lista_nacional) dados_agr <- lapply(lista_agr, bind_rows) # ============================================================================= # 3) RÓTULOS E CORES # ============================================================================= labels_decis <- c( "Decile 1: bottom 10%", "Decile 2: 10–20%", "Decile 3: 20–30%", "Decile 4: 30–40%", "Decile 5: 40–50%", "Decile 6: 50–60%", "Decile 7: 60–70%", "Decile 8: 70–80%", "Decile 9: 80–90%", "Decile 10: top 10%" ) cores <- colorRampPalette(c("#D7191C", "#FFFFBF", "#2B83BA"))(10) # ============================================================================= # 4) CONSTRUÇÃO DO GRÁFICO # ============================================================================= fig_220_3 <- plot_ly() # ── 4A) DECIS INDIVIDUAIS ───────────────────────────────────────────────────── # CHAVE: SEM legendgroup compartilhado → cada trace é independente na legenda for (i in 1:10) { df_d <- dados_decis %>% filter(decil_num == i) fig_220_3 <- fig_220_3 %>% add_trace( data = df_d, x = ~ano, y = ~media, type = 'scatter', mode = 'lines+markers', name = labels_decis[i], line = list(color = cores[i], width = 2.5), marker = list(color = cores[i], size = 7), error_y = list( type = "data", array = ~upper - media, arrayminus = ~media - lower, visible = TRUE, thickness = 1, width = 2, color = cores[i] ), hoverinfo = "text", text = ~paste0( "<b>Year:</b> ", ano, "<br>", "<b>", labels_decis[i], "</b><br>", "<b>Mean Income:</b> R$ ", formatC(media, format = "f", digits = 0, big.mark = ","), "<br>", "<b>CI (95%):</b> [R$ ", formatC(lower, format = "f", digits = 0, big.mark = ","), " – R$ ", formatC(upper, format = "f", digits = 0, big.mark = ","), "]" ), # ↓ legendgroup ÚNICO por decil → toggle individual legendgroup = paste0("decil_", i), showlegend = TRUE ) } # ── 4B) GRUPOS AGREGADOS (linha tracejada, marcador triângulo) ──────────────── for (g in seq_along(GRUPOS_AGREGADOS)) { grp <- GRUPOS_AGREGADOS[[g]] df_g <- dados_agr[[g]] fig_220_3 <- fig_220_3 %>% add_trace( data = df_g, x = ~ano, y = ~media, type = 'scatter', mode = 'lines+markers', name = grp$nome, line = list(color = grp$cor, width = 3, dash = grp$dash), marker = list(color = grp$cor, size = 9, symbol = "triangle-up", line = list(color = "white", width = 1)), error_y = list( type = "data", array = ~upper - media, arrayminus = ~media - lower, visible = TRUE, thickness = 1.5, width = 3, color = grp$cor ), hoverinfo = "text", text = ~paste0( "<b>Year:</b> ", ano, "<br>", "<b>", grp$nome, "</b><br>", "<b>Mean Income:</b> R$ ", formatC(media, format = "f", digits = 0, big.mark = ","), "<br>", "<b>CI (95%):</b> [R$ ", formatC(lower, format = "f", digits = 0, big.mark = ","), " – R$ ", formatC(upper, format = "f", digits = 0, big.mark = ","), "]" ), legendgroup = paste0("agregado_", g), showlegend = TRUE ) } # ── 4C) MÉDIA NACIONAL (camada superior) ────────────────────────────────────── fig_220_3 <- fig_220_3 %>% add_trace( data = dados_nacional, x = ~ano, y = ~media, type = 'scatter', mode = 'lines+markers', name = "<b>National Average</b>", line = list(color = COR_MEDIA_NACIONAL, width = LARGURA_MEDIA_NACIONAL), marker = list(color = COR_MEDIA_NACIONAL, size = TAMANHO_MARCADOR_MEDIA, symbol = "diamond", line = list(color = "white", width = 1.5)), error_y = list( type = "data", array = ~upper - media, arrayminus = ~media - lower, visible = MOSTRAR_IC_MEDIA_NACIONAL, thickness = 2, width = 3, color = COR_MEDIA_NACIONAL ), hoverinfo = "text", text = ~paste0( "<b>Year:</b> ", ano, "<br>", "<b>★ NATIONAL AVERAGE ★</b><br>", "<b>Mean Income:</b> R$ ", formatC(media, format = "f", digits = 0, big.mark = ","), "<br>", "<b>CI (95%):</b> [R$ ", formatC(lower, format = "f", digits = 0, big.mark = ","), " – R$ ", formatC(upper, format = "f", digits = 0, big.mark = ","), "]" ), legendgroup = "nacional", showlegend = TRUE ) # ============================================================================= # 5) LAYOUT # ============================================================================= fig_220_3 <- fig_220_3 %>% layout( title = list( text = "<b>Mean per capita household income by income decile in Brazil, 1992–2022</b>", x = 0.5, y = 0.97, font = list(size = 20) ), xaxis = list( title = "<b>Year</b>", gridcolor = "#E5E5E5", tickmode = "array", tickvals = anos_disponiveis, tickangle = -45, font = list(size = 10) ), yaxis = list( title = "<b>Mean per capita household income (BRL)</b>", tickprefix = "R$ ", gridcolor = "#E5E5E5" ), legend = list( x = 0.01, y = 0.99, xanchor = "left", yanchor = "top", bgcolor = 'rgba(255,255,255,0.85)', bordercolor = "#CCCCCC", borderwidth = 1, font = list(size = 9), traceorder = "reversed" ), annotations = list(list( x = 0.5, y = -0.15, text = "Source: Prepared by the author based on PNAD/PNAD-C (IBGE) data, harmonized by Salata et al. (2025).<br>Article, code and data available at doi:10.7910/DVN/UZ2SHW", showarrow = FALSE, xref = 'paper', yref = 'paper', xanchor = 'center', yanchor = 'top', font = list(size = 14, color = "#555555") )), margin = list(t = 90, b = 120, l = 80, r = 40), plot_bgcolor = "white", paper_bgcolor = "white", hovermode = "closest" ) fig_220_3 cat("\n✅ Gráfico 220_3 gerado!\n") cat("\nMELHORIAS IMPLEMENTADAS:\n") cat("━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n") cat("✓ Toggle INDIVIDUAL por decil (legendgroup único por trace)\n") cat("✓ Bottom 40% (D1–D4) → rosa, linha tracejada, triângulo\n") cat("✓ Bottom 50% (D1–D5) → laranja, dash-dot, triângulo\n") cat("✓ Bottom 80% (D1–D8) → roxo, dotted, triângulo\n") cat("✓ Agregados calculados com svymean() no subconjunto dos decis\n")

This project applies unsupervised learning to a dataset of League of Legends champion base statistics sourced from Kaggle (Cute Dango, League of Legends Champions dataset, available at kaggle.com/datasets/cutedango/league-of-legends-champions) to discover whether champions naturally cluster into distinct statistical archetypes, and which features drive those groupings. The workflow combines three complementary approaches: Hard clustering (K-Means, Hierarchical) to identify stable, discrete champion archetypes, Soft clustering (Fuzzy C-Means) to quantify champion hybridity, how strongly each champion belongs to one archetype versus another, Dimensionality reduction (PCA, MDS, UMAP, t-SNE, SOM) to visualize the structure of the feature space and validate clustering results across multiple independent methods.

This is the completion of 6th lab for class Epi 553.

This project analyzes a complete archive of 90,000+ posts made by Donald Trump on X (Twitter) and Truth Social from 2009–2026. Using R, Python, and Bash, I clean, filter, and structure the data to focus exclusively on original, text-based posts—excluding reposts, quotes, links, images, and videos—to isolate Trump’s direct public communication. The objective is to examine rhetorical patterns and topic trends over time through large-scale text classification. I analyze the frequency of themes such as religion, immigration, education, and economics; measure sentiment toward groups defined by ethnicity, religion, and sexual orientation; track posting behavior; and investigate which linguistic and structural features are associated with viral posts. Overall, the project integrates data wrangling, analysis, and computational text modeling to study the evolution and impact of modern political discourse.

DATA 607 Assignment 5B Elo expected score calculation using cleaned up chess tournament data set clean from a text file in Project 1

DATA 607 Assignment 5A analysis