· EXPLORE BLACK INVENTIONS LIVE 1870–1940
·
BLACK CREATIVE DATA EXPLORER

The Golden Era of Black Invention

Census-linked patent records from the Golden Age of American Invention. Explore all ~50,000 Black American patents, 1870–1940.

DATA: ICPSR 120556 / 109970 / 193421 · BROOKINGS 2020 · COOK 2014 · LIVE —

PATENT ACTIVITY
1870–1940
NORTH vs SOUTH
RELATIVE ODDS
● NORTH — 2.10× odds ● SOUTH — 0.30× odds
Black Northerners were 8× more likely to patent than Southerners. Jim Crow, racial violence, and exclusion from patent resources created measurable suppression. Cook estimates 1,100+ missing patents from violence alone. With 50,000 total patents, Black inventors outpaced immigrants from every country except England and Germany.
DATA SOURCES
7 DATASETS
50K METHODOLOGY: Census-matched data identifies 2.7% of patents as Black-invented. The 5.4× multiplier accounts for unmatched records. Extrapolating across all 71 years yields ~49,800 total Black patents. 87% of all inventions were by US-born Americans.
SOURCES & METHODOLOGY
REFERENCES

PRIMARY Sarada, Andrews & Ziebarth (2019). "Changes in the Demographics of American Inventors." Explorations in Econ. History. ICPSR 109970.

PRIMARY Andrews & Rothwell (2023). "Reassessing the Contributions of Black Inventors." Essays in Econ. & Business History. ICPSR 193421.

ESTIMATE Rothwell, Perry & Andrews (2020). "The Black Innovators." Brookings Institution. Source of the 50,000 estimate.

VIOLENCE Cook (2014). "Violence and Inventive Activity." J. of Econ. Growth. 726 confirmed patents; 1,100+ missing.

ARCHIVAL Baker (1902/1913). Patent Office surveys. 400+ → ~800 Black patents identified.

EXPLORE BLACK INVENTIONS
50 RECORDS
YEAR –
FULL PATENT DATABASE 1,992,169 RECORDS · 1870–1940
SEARCH TO EXPLORE
YEAR –
Search across all 1,992,169 patent records
Try: hair comb · locomotive · ventilator · Jones · Chicago
BLK:INV · EXPLORE BLACK INVENTIONS
DATA: ICPSR 120556 / 109970 / 193421 · BROOKINGS 2020 · COOK 2014
GEOGRAPHIC DISTRIBUTION OF BLACK INVENTORS
INTERACTIVE MAP
Hover over any state to see Black inventor counts. Color intensity reflects density relative to total patents.
0
120+
Color Scale: ■ Deep blue = few or no Black inventors identified   ■ Green = moderate presence (5–30)   ■ Amber = significant hub (30–50)   ■ Red/Pink = major center (50+). Colors represent census-confirmed Black inventors 1870–1940. Click any state to filter the Patent Explorer below or read about Black innovation in that state.
The industrial North (NY, PA, OH, IL, MA) dominates Black patenting. Washington D.C. punches far above its size due to Howard University and the federal patent office. Jim Crow laws, racial violence, and exclusion from resources suppressed Southern innovation.
BY THE NUMBERS: THE 50,000 STORY
METHODOLOGY
~50,000
ESTIMATED BLACK PATENTS
1870-1940 Extrapolated from census-match rate
1,992,169
TOTAL US PATENTS IN DATABASE
Parsed from Annual Reports of the Commissioner of Patents
27,120
CENSUS-MATCHED INVENTORS
Linked to decennial census records 1870-1940
755
CONFIRMED BLACK INVENTORS
2.78% of census-matched aligning with Rothwell-Andrews
1,100+
MISSING DUE TO VIOLENCE
Cook (2014) patents lost to lynchings and race riots
8x
NORTH vs SOUTH GAP
Black Northerners 8x more likely to patent than Southerners
How do we get to 50,000? Census-matched data identifies 2.78% of patentees as Black-invented. The census only captures ~64% of inventors in any given decade. Applying the 2.78% rate across all 1.8 million utility patents filed between 1870 and 1940 yields approximately 49,800 total Black patents. 87% of all inventors were US-born Americans. The 5.4x multiplier accounts for unmatched records, non-census years, and inventors who could not be linked to census data due to name variations or migration.

This is consistent with the Brookings Institution estimate (Rothwell, Perry and Andrews, 2020) and the econometric models of Sarada, Andrews and Ziebarth (2019). With 50,000 patents, Black inventors produced more innovations than immigrants from every country except England and Germany. Sources: ICPSR 109970, ICPSR 193421, Brookings 2020
ESTIMATED BLACK PATENTS BY DECADE
CULTURAL PROPERTY VALUATION
CPVF ENGINE
MADECX Cultural Property Valuation Framework™ — Each patent is scored across five dimensions: Cultural Influence (CIS), Consumer Conversion (CCI), Likeness & Identity (LIP), Commercial Usage (CUV), and Heritage & Lineage Multiplier (HLM). Scores produce a Total Creative Property Market Value (TCPMV) with tier classification and monetary range.
Valuations computed from patent metadata, census data, and historical research. Enrichment from USPTO citations, Wikipedia, and museum holdings will increase confidence.
PORTFOLIO TIER DISTRIBUTION Based on CPVF score thresholds
TOP VALUED PATENTS
LOST COMMERCIAL VALUE — THE WEALTH GAP
LCV ENGINE
What would Black inventors have accumulated? The Lost Commercial Value calculates what inventors and their descendants would have earned if they had received standard royalties and reinvested at market rates from patent date to present. Three layers: Lost Royalties (industry revenue × attribution × royalty rate), Compounded Wealth (royalties grown at 4–10%), and Derivative Participation (share of industries built on the invention). This is the quantified cost of exclusion.
MODERATED DISCUSSION
3 APPROVED
🛡 MODERATION ACTIVE — Evidence-based discussion. MOD = verified moderator.
REVIEWED BEFORE PUBLICATION
BLK:INV EXPLORE BLACK INVENTIONS
NAVIGATION
DATASET STATS
EST. PATENTS~50,000
PERIOD1870–1940
CONFIRMED (COOK)726
N vs S GAP8×
SHARE US PATENTS2.7%
DB STATUS● LIVE
DATABASE TOOLS
DATA SOURCES
BLK:INV · EXPLORE BLACK INVENTIONS
Powered by Supabase · ICPSR 109970
MADE CX · Cultural Property Valuation Framework
DATABASE CHARACTER AUDIT
Mojibake investigation · Stata .dta → CSV → Postgres encoding pipeline

● ENCODING SCAN RESULTS

—
CORRUPTED RECORDS FOUND
—
CLEAN RECORDS
—
TOTAL SCANNED
—
% AFFECTED

● CORRUPTED RECORD SAMPLES

RAW NAMECLEANEDSTATEPATTERN
Run scan to populate…

● ENCODING ROOT CAUSE

The corruption originates in the ICPSR Stata .dta export pipeline. Middle initials and special characters (periods, apostrophes) were encoded as UTF-8 multi-byte sequences, then those raw bytes were re-imported into Postgres as if they were Latin-1/Windows-1252, producing double-encoded mojibake.
RAW SEQUENCEHEX (Latin-1)LIKELY ORIGINALEXAMPLE
ã¢ÂÂe3 a2 c2 c2. (period)Frank Bã¢Â ROSS → Frank B ROSS
ã·e3 c2 b7. (period)Henry Lã· BROWNBACK → Henry L BROWNBACK
ãÂe3 c2(truncated letter)Samuã → Samuel (el lost)
ÃÂc3 c2(lost letter A–Z)Theodore ÃÂ MILLER → A. MILLER
ÃÂà Âc3 c2 ...(multi-char loss)WilliãÂà→ William (am lost)

● SUPABASE SQL REMEDIATION

Run these queries in your Supabase SQL Editor to clean the database permanently.
STEP 1 — AUDIT COUNT
SELECT COUNT(*) AS corrupted_count FROM inventors WHERE first_name ~ '[^\x00-\x7F]' OR last_name ~ '[^\x00-\x7F]';
STEP 2 — PREVIEW BEFORE CLEANING
SELECT id, first_name, last_name, state, regexp_replace(first_name, '[^\x00-\x7F]+', '', 'g') AS clean_first, regexp_replace(last_name, '[^\x00-\x7F]+', '', 'g') AS clean_last FROM inventors WHERE first_name ~ '[^\x00-\x7F]' OR last_name ~ '[^\x00-\x7F]' ORDER BY last_name LIMIT 50;
STEP 3 — APPLY FIX (irreversible — backup first)
UPDATE inventors SET first_name = regexp_replace(first_name, '[^\x00-\x7F]+', '', 'g'), last_name = regexp_replace(last_name, '[^\x00-\x7F]+', '', 'g') WHERE first_name ~ '[^\x00-\x7F]' OR last_name ~ '[^\x00-\x7F]'; -- Returns: UPDATE N (affected row count)
⚠ The frontend already applies cleanName() to strip mojibake on display. The SQL fix makes the cleaning permanent in the database.