tidycops

tidycops is an incident-focused R package for pulling public police data across cities with one interface.

It supports two core modes:

• Comparable cross-city output (std_* schema)
• Full city-native output (all exposed source fields)

Currently Supported Cities

As of April 30, 2026, incident adapters are available for:

• Boston
• Chicago
• Cincinnati
• Cleveland
• Dallas
• Denver
• Detroit
• Fort Lauderdale
• Gainesville
• Grand Rapids
• Hartford
• Houston
• Indianapolis
• Kansas City
• Minneapolis
• Naperville
• New Orleans
• New York City
• Pittsburgh
• Providence
• Rochester
• San Antonio
• San Francisco
• Seattle
• Washington, DC

Run list_supported_incident_cities() for the latest in-package list.

Installation

You can install the development version of tidycops like so:

# install.packages("remotes")
remotes::install_github("Steal-This-Code/tidycops")

Incident Workflow

library(dplyr)

# Which cities are wired?
list_supported_incident_cities()

# Which fields are most comparable across cities?
list_common_incident_fields(min_cities = 4)

# What fields (standardized and source) are available for a specific city?
list_incident_fields("chicago")
list_incident_fields("san_francisco")

1) Comparable Cross-City Output

Use this when you want harmonized fields for city-to-city analysis.

chicago_compare <- get_incidents(
  city = "chicago",
  view = "comparable",
  start_date = "2026-04-01",
  end_date = "2026-04-07",
  as_sf = TRUE,
  limit = 500
)

2) Full City Schema Output

Use this when you want all city-native fields plus adapter metadata.

cincy_full <- get_incidents(
  city = "cincy",
  view = "city_full",
  start_date = "2026-04-01",
  end_date = "2026-04-07",
  limit = 1000
)

3) Raw Payload Output

Use this when you want untouched source payload rows only.

cincy_raw <- get_incidents(
  city = "cincy",
  view = "city_raw",
  start_date = "2026-04-01",
  end_date = "2026-04-07",
  limit = 1000
)

# Equivalent convenience helper:
cincy_raw_2 <- download_city_incidents_raw(
  city = "cincy",
  start_date = "2026-04-01",
  end_date = "2026-04-07",
  limit = 1000
)

Date Shortcut

Use last_n_days when you want a trailing window without manual date math.

recent_cincy <- get_incidents(
  city = "cincinnati",
  view = "comparable",
  last_n_days = 14,
  limit = 1000
)

Source Freshness and Bounded Feeds

Some adapters are rolling windows, capped historical feeds, or scope-limited (for example calls-for-service or Part I only). Check source metadata before analysis.

list_incident_sources() |>
  dplyr::filter(source_status %in% c("historical_capped", "rolling_window") | scope_status != "all_incidents")

Build Priorities: Incident Function Shortcuts

Before expanding to arrests/UOF, these are the highest-value additions for incident workflows:

1. dedupe controls
   • Add optional dedupe modes (none, incident_id, incident_number) for city feeds that publish multiple offense rows per incident.
2. fields = "core" projection
   • Add a quick column subset option for the most cross-city stable fields.
3. Source diagnostics helper
   • Add a helper to report expected coverage, null rates, and key field completeness by city/date window.

These keep the package incident-first while making pull/analyze loops faster.