Data
4 unrolled threads about data, each one readable on a single page.
- Nov 19, 2025
White by Default? The viral posts were right. We scraped 5.5 million criminal records and 1.5 million mugshots from 39 states. 29% of Hispanics are being misclassified as White in official Departme...
SummaryResearchers analyzed 5.5 million criminal records and 1.5 million mugshots from 39 states, finding that approximately 29% of Hispanics are systematically misclassified as White in official Department of Corrections databases, even when Hispanic classification is explicitly available. Using machine learning trained on facial recognition and name demographics, they achieved 92.76% accuracy in racial classification and confirmed the misclassification bias is real, not anecdotal—correcting for it increases reported Hispanic criminal rates by 20-31% while decreasing White rates by 4-6%.11 tweets1 min read - Sep 5, 2025
1/ 📅 On September 5 at 12:30 PM (UTC), The BLS will release the Non-Farm Payrolls (NFP) report — the most important monthly indicator of labor market health. Expectations: +75K jobs (prior 73K) and...
SummaryThe BLS will release the Non-Farm Payrolls report on September 5 at 12:30 PM UTC, with expectations of +75K jobs and 4.3% unemployment. This is the most important monthly labor market indicator.2 tweets1 min read - Jun 8, 2026
Every legitimate vote in LA deserves to count, but a broken system dilutes our community's voice. Latinos Por Pratt is launching a data-driven election audit of the June 2 mayoral primary. We aren't w...
SummaryLatinos Por Pratt is launching a data-driven audit of LA's June 2 mayoral primary, using voter file analysis and crowdsourced field verification to identify fraudulent addresses and build a legal case to contest the election results before the July 7 filing deadline.4 tweets1 min read - Nov 13, 2025
If you’re still sending raw JSON into your LLMs, you’re burning tokens, latency, and budget! Try TOON (Token-Oriented Object Notation). Clear like YAML, compact like CSV: • 30–60% fewer tokens • Up...
SummaryTOON (Token-Oriented Object Notation) is a new data format that reduces token usage by 30-60% compared to JSON, lowering costs and latency for LLM applications. The thread introduces the open-source tool, benchmarking resources, and conversion utilities to help developers optimize their data formatting.8 tweets1 min read