Open Census MCP Server

# Census Data Retrieval Script # Called by Python MCP server to fetch ACS data via tidycensus library(tidycensus) library(dplyr) library(jsonlite) # Parse command line arguments args <- commandArgs(trailingOnly = TRUE) if (length(args) < 4) { stop("Usage: Rscript census_data_retrieval.R <location> <variables> <year> <survey>") } location_json <- args[1] variables_json <- args[2] year <- as.numeric(args[3]) survey <- args[4] # Parse JSON inputs location_data <- fromJSON(location_json) variables_list <- fromJSON(variables_json) # Set Census API key if available census_api_key <- Sys.getenv("CENSUS_API_KEY") if (nchar(census_api_key) > 0) { census_api_key(census_api_key) } # Main data retrieval function get_census_data <- function(location_data, variables_list, year, survey) { tryCatch({ # Determine geography type and codes geography <- location_data$geography state <- location_data$state county <- location_data$county place <- location_data$place # Call get_acs based on geography if (geography == "state") { data <- get_acs( geography = "state", variables = variables_list, year = year, survey = survey, state = state ) } else if (geography == "county") { data <- get_acs( geography = "county", variables = variables_list, year = year, survey = survey, state = state, county = county ) } else if (geography == "place") { # Get all places in state, then filter for specific place data <- get_acs( geography = "place", variables = variables_list, year = year, survey = survey, state = state ) # Filter for the specific place if specified if (!is.null(place) && nchar(place) > 0) { # Create search patterns for place matching place_patterns <- c( paste0("^", place, "$"), # Exact match paste0("^", place, ","), # Place followed by comma paste0(place, " city,"), # City suffix paste0(place, " town,"), # Town suffix paste0(place, " village,") # Village suffix ) # Try each pattern until we find a match filtered_data <- NULL for (pattern in place_patterns) { filtered_data <- data[grepl(pattern, data$NAME, ignore.case = TRUE), ] if (nrow(filtered_data) > 0) { break } } # If no match found, try partial matching if (is.null(filtered_data) || nrow(filtered_data) == 0) { # Extract base place name for partial matching base_place <- gsub(" (city|town|village)$", "", place, ignore.case = TRUE) filtered_data <- data[grepl(base_place, data$NAME, ignore.case = TRUE), ] } # Use filtered data if found, otherwise return error if (!is.null(filtered_data) && nrow(filtered_data) > 0) { data <- filtered_data } else { stop(paste("No data found for place:", place, "in state:", state)) } } } else if (geography == "us") { data <- get_acs( geography = "us", variables = variables_list, year = year, survey = survey ) } else { stop(paste("Unsupported geography:", geography)) } # Format output result <- list( data = data, source = paste("US Census Bureau American Community Survey", survey, "Estimates"), year = year, survey = toupper(survey), geography = geography, success = TRUE ) # Convert to JSON and print cat(toJSON(result, auto_unbox = TRUE, pretty = TRUE)) }, error = function(e) { error_result <- list( error = as.character(e), success = FALSE ) cat(toJSON(error_result, auto_unbox = TRUE)) }) } # Execute main function get_census_data(location_data, variables_list, year, survey)