Geography

Where the archives say it happened

Two licensed historical datasets of US shootings, mapped and tabulated with their own definitions attached. They disagree about what counts, they stopped collecting years ago, and neither one is a census of the country.

Not current coverage. The CNN school archive ends in 2018 and the Stanford mass-shootings archive ends in 2016. Nothing here describes what has happened since. An empty result for a recent date range means no archive covers it, not that no shooting occurred.

Sources, licences and archive health

State: archive. Historical archives; not current US shooting coverage.

CNN school shootings archive

Coverage 2009-01-08 to 2018-12-11 · not current coverage · Archived; no scheduled refresh

What this source counts: K–12 school-property incidents in which at least one person other than the shooter was shot; includes qualifying accidental discharges and BB-gun injuries, with the stated law-enforcement/security exception.

Source rows: 180
Included: 180
Flagged uncertain: 0
Excluded by the source: 0
Rows with no coordinate: 3

What this source says about its own limits

Historical selected incidents, not a current or exhaustive US census.
Coordinates are supplied by the source; point precision is unknown.
County and metropolitan-area assignments are unavailable.

Stanford Mass Shootings in America archive

Coverage 1966-08-01 to 2016-06-26 · not current coverage · Archived; no scheduled refresh

What this source counts: At least three shooting victims, excluding the shooter; excludes identifiably gang-, drug-, or organized-crime-related shootings. Included, uncertain, and excluded source flags are retained.

Source rows: 335
Included: 325
Flagged uncertain: 6
Excluded by the source: 4
Rows with no coordinate: 0

Stanford Mass Shootings in America, courtesy of the Stanford Geospatial Center and Stanford Libraries. Licensed CC-BY-4.0. Pinned to commit 7025a75fdc (2018-03-26), 489,585 bytes, SHA-256 9d42ea9d72bc7280…, checked 2026-09-13. No endorsement by the source is implied.

What this source says about its own limits

Historical selected incidents, not a current or exhaustive US census.
Coordinates are supplied by the source; point precision is unknown.
County and metropolitan-area assignments are unavailable.
The source warns that changing online reporting creates historical ascertainment bias.
Native CaseIDs 156 and 168 are reused for different incidents; explicit disambiguation preserves all rows.

515 source rows across both archives reconcile to 1 incidents in this selection, of which 1 carry a coordinate and 0 do not. 4 cross-source overlaps were manually reviewed and merged.

Filter

State

DefinitionThe two definitions are different measurements, not two slices of one.

Date from

Date toInclusive. Archive coverage ends in 2018.

Include records the source flagged uncertainStanford flags six records as uncertain rather than dropping them. They are excluded by default and marked wherever they appear.

Clear filters

What this selection contains

Incidents matching: 1
On the map: 1
Table only, no coordinate: 0
School definition: 1
Mass definition: 0
Date unknown: 0

Showing in Florida, dated 2011-04-13. The two definition counts overlap and must not be added: 4 reviewed events appear under both. Uncertain records are excluded in this selection.

Map of 1 located incidents

Skip the map’s 1 points and go to the incident table

School — CNN archive, 2009 to 2018. K–12 school property, at least one person other than the shooter shot.
Mass — Stanford archive, 1966 to 2016. At least three shooting victims excluding the shooter.
In both archives — one of the four manually reviewed overlaps. The two definitions are kept separate; the counts are not added.
Hollow marker — the source flagged the record uncertain. Shown only while “include uncertain records” is on.

1 mapped point, every incident this filter matched that carries a coordinate — not the thirty rows below. Dates run from 2011-04-13 to 2011-04-13.

Albers equal-area conic projection of the coordinates the archives supply, scaled to fit the current selection. This map displays incident points, without boundary outlines or shaded regions. The pinned Census county and metro polygons are used separately to derive the approximate place labels below; they are modern boundaries, not event-date boundaries. Point precision is unknown: these are the source files’ own coordinates, not verified street addresses, and markers overlap where incidents are close together. Drawn extent 28.7°N to 28.7°N, 81.5°W to 81.5°W.

Selecting a point filters the table to that incident’s date and state and jumps to its row.

Counts by place

Counts of archived incidents per state, largest first. These are counts, not rates, and not a ranking of risk: the archives select for particular kinds of event and stop at different years.

Florida (FL)1

County and metropolitan area

Approximate, and not the boundaries of the day. January 1, 2026 boundaries applied to historical incident points; not event-date boundaries.

Each label is produced by testing the source archive’s own coordinate against a pinned boundary polygon. The coordinate’s precision is unknown and is not a verified street address, the polygons are generalized to 0.001 degrees, and a point within 0.002 degrees of an edge is left unknown rather than assigned. An incident’s county label can therefore name an area that did not exist under that name on the date of the incident.

Derived from U.S. Census Bureau TIGER geographic data; not endorsed by the Census Bureau. Boundaries: U.S. Census Bureau TIGERweb, 2026-01-01 vintage, retrieved 2026-09-13. US federal government work (17 USC 105).

Which boundary layers, and how they were checked

Counties (or statistically equivalent entities); January 1, 2026 vintage3,235 features, independently counted at 3,235; 8,147,620 bytes, SHA-256 4445f6eb4d509e66…
Metropolitan Statistical Areas; January 1, 2026 vintage393 features, independently counted at 393; 1,893,130 bytes, SHA-256 055b5ef53c15657e…

Method ray_cast_polygon_holes_v1; event-date boundaries: no. Licence: US federal government work (17 USC 105) (reference). Projection version 2c96e789c4fc6df1….

Counties and county equivalents

1 of 1 incidents in this selection carry a derived label, across 1 area. These counts cover the whole selection, not the table page below.

Orange County, FL (12095)1

0 incidents have no county or county equivalent label. That is a refusal to guess, not a count of zero, and those incidents stay in the table and on the map. The reasons an incident lands here:

near_boundary: the source point falls within the boundary tolerance, so no area is claimed rather than one being picked
overlap: the source point matched more than one area, so neither is claimed
outside: the source point falls outside every boundary in the layer
source_state_conflict: the derived area's state disagrees with the state the source row states, so neither is claimed
missing_point: the source row carries no coordinate at all
invalid_point: the source row's coordinate is not a usable position

Metropolitan areas

1 of 1 incidents in this selection carry a derived label, across 1 area. These counts cover the whole selection, not the table page below.

Orlando-Kissimmee-Sanford, FL Metro Area (36740)1

0 incidents have no metropolitan area label. That is a refusal to guess, not a count of zero, and those incidents stay in the table and on the map. The reasons an incident lands here:

county_near_boundary: the county lookup was inconclusive (the point is within the boundary tolerance), so no metro is claimed either
county_overlap: the county lookup matched more than one area, so no metro is claimed either
county_outside: the point falls outside every county, so no metro is claimed either
missing_point: the source row carries no coordinate at all
county_invalid_point: the coordinate is not usable, so no metro is claimed

Still no rates. Counts by county or metro area are not rates by county or metro area. No population denominator compatible with these archives’ selection criteria is supplied, and dividing by one from elsewhere would compare a selected, definition-specific sample against a complete population.

Rates per population: unavailable

Counts are not rates. No compatible population denominator is supplied with these archives, and dividing archive counts by a population figure from elsewhere would compare a selected, definition-specific sample against a complete one. Until a matching denominator exists, this page shows counts only.

Incidents

Incidents in Florida, dated 2011-04-13. Page 1 of 1, 30 rows per page, 1 matching in total.
Date	Incident	Place	Definition	Location on map	Source rows
2011-04-13	Sheeler Charter High School	Apopka, FloridaCounty: Orange County, FL (approximate, 2026 boundary)Metro: Orlando-Kissimmee-Sanford, FL Metro Area (approximate, 2026 boundary)	School (CNN definition)	28.6732, -81.4941Precision unknown; supplied by CNN school shootings archive.	CNN school shootings archiverow 2985474fd2cb3fb157208bfc, as written: 13-Apr-11

Overlaps between the two archives

Only explicit reviewed named-venue links merge source records. Other overlaps may remain; category counts are not additive.

Chardon High School · Chardon High School
Manually reviewed date, state and named-school agreement in both pinned datasets.
source row 1 · source row 2
Sandy Hook Elementary · Sandy Hook Elementary School
Manually reviewed date, state and named-school agreement in both pinned datasets.
source row 1 · source row 2
Sparks Middle School · Sparks Middle School
Manually reviewed date, state and named-school agreement in both pinned datasets.
source row 1 · source row 2
Marysville-Pilchuck High School · Marysville-Pilchuck High School
Manually reviewed date, state and named-school agreement in both pinned datasets.
source row 1 · source row 2
2016-02-09 · unresolved candidate, counted separately
Muskegon Heights: date/city and generic high-school description suggest overlap, but named-venue corroboration is insufficient; not merged. Stanford marks it uncertain.

Limits of this data

Historical selected incidents; reporting and definition differences prevent treating these as current national rates.
Only explicitly reviewed cross-source overlaps are merged; other overlaps may remain. Category counts are not additive.
Source coordinates have unknown precision and are not verified rooftop locations.
County and metro labels are approximate point-in-polygon estimates using January 1, 2026 Census boundaries, not event-date boundaries.
Near-boundary, ambiguous and missing points remain unknown; non-metropolitan is separate from unknown.
Rates are disabled because no compatible population denominator is supplied.

Archive version fed61bae40f2764e…, imported 2026-09-13 04:08:57 UTC, rendered 2026-09-13 12:33:41 UTC.