Baselight

Finding it hard to analyze data with SQL? That changes today with Baselight AI, your data analyst copilot! Baselight AI lets you extract insights from Baselight datasets (and your own uploads!), audit reasoning, and verify the data used to reach conclusions. We're launching today on @ProductHunt!

Link to Baselight Daily Insights: baselight.app/daily-insights Bookmark it and check it every morning - or add it to your agent's feed so it can stay on top of what matters.

Baselight update: 240M rows of U.S. mortgage and housing data, +4B total rows, and the launch of Baselight Daily Insights. This week we added Federal Housing Finance Agency data across Federal Home Loan Bank and Fannie Mae / Freddie Mac public-use datasets. That means almost 2 decades of U.S. mortgage and housing data: borrower demographics, credit profiles, loan characteristics, census tract geography, single-family loans, and multifamily properties. You can now explore questions like: - Are young people being priced out? - Where are apartments being built? - Are there lending disparities by race and ethnicity? Baselight now has: 480,465,801,488 rows (+4B this week) 459,012 tables (+1K) 70,948 datasets (+251) We also launched Baselight Daily insights generated by agents that find anomalies, cross-check evidence, rank what matters, and publish the query trail. That means every insight is grounded in structured data, with the underlying queries available to inspect instead of asking you to trust a black-box AI summary. Bookmark it and check it every morning (link in the comments) - or add the feed to your agent so it knows what matters every day.

Baselight Pulse / Daily Insights: baselight.app/daily-insights/

Baselight Pulse is live. Every day, our agent team scans Baselight’s structured data and crafts a fresh set of Daily Insights across: - Soccer - Crypto prices - Stocks The goal is simple: move beyond search and dashboards into proactive intelligence. Instead of waiting for users to ask the right question, Baselight Pulse starts surfacing what may matter: anomalies, divergences, inflection points, unusual patterns, and market or sports moments worth paying attention to. And because these insights are generated from structured data, every insight includes the exact evidence trail behind it: the queries, intermediate results, and data used to reach the conclusion. So this is not just "AI says so." It is auditable, inspectable, and explainable intelligence. We’ve been preparing this over the last couple of weeks, and it feels like another important step towards the kind of data agents we want to build: agents that understand large-scale structured data, detect what changed, and help users discover insights they might not have known to look for. This is still early, and we’d love feedback. Link in the comments. Signal, not noise.

Query it live on Baselight: baselight.ai

Another big week for Baselight. We’ve added the WHO Global Health Observatory: 10.3 million rows of data, across 3,057 indicators, spanning 1932–2030. You can now explore questions like: - How large is the global life expectancy gap? - How does tobacco use prevalence differ by sex worldwide? - What share of TB patients test HIV-positive? Baselight now stands at: 476,066,264,894 rows (+3B) 457,900 tables (+2K) 70,697 datasets (+1K) Approaching half a trillion rows of structured data, all in one place.

Query it live on Baselight: baselight.ai

Most blockchain data is technically public. Very little of it is actually ready to use. This week we launched tempo_decoded on Baselight - a structured dataset that turns Tempo’s raw on-chain event logs into clean, queryable, AI/analytics-ready tables. Tempo is a purpose-built Layer 1 for payments, designed for high-throughput, low-cost global transactions, including machine payments. But raw blockchain logs are still messy, fragmented, and difficult to analyze quickly. So we did the hard part. In just the last few days of data, tempo_decoded already includes: - 902,000+ decoded events from 1,876+ unique smart contracts - 594,619 Transfer events - 60,663 Approval events - 16,104 Swap events across 70 DEX contracts - 8,547 NFT listings + 4,701 NFT sales - 2,989 Stake events + 2,482 Unstake events - 1,316 bridged withdrawals This is the kind of infrastructure we believe data teams and AI systems need more of: not just access to raw data, but access to data that’s already structured, interpretable, and ready to query. Baselight now stands at: 472,946,369,391 rows (+2B this week) 455,713 tables (+338) 69,954 datasets (+33) We’re excited to keep expanding Baselight’s coverage across domains.

Links to the new datasets: - Baseball: baselight.app/u/blt/dataset/… - UFC / MMA: baselight.app/u/blt/dataset/…

Baselight is getting closer to full circle for sports and prediction market analysis. We already have Kalshi, Polymarket, and other prediction market data in the platform. Now we’ve added two more datasets: - Baseball: global baseball data covering competitions, teams, games with inning-by-inning scoring, standings, season stats, and pre-game odds - UFC / MMA: covering events, fighters, training camps, fights, results, per-fighter fight stats, and pre-fight odds That builds on the sports coverage already in Baselight across soccer, basketball, NFL, and Formula 1 - and makes the overall surface much stronger for cross-market analysis, forecasting, and finding pricing inefficiencies. More sports data. More market context. More ways to connect the dots.

More sports. More onchain data. More surface area for AI. Next week, Baselight adds the Ultimate Baseball Dataset, the Ultimate MMA / UFC Dataset, and Tempo blockchain data. That means structured coverage of global baseball leagues, deep UFC data across fighters and fights, and easier analysis on top of Tempo - a purpose-built Layer 1 for payments, built with leading fintechs and Fortune 500s to support high-throughput, low-cost global transactions, including machine payments. Baselight now stands at: - 470,690,344,848 rows (+5B this week) - 455,375 tables (+185) - 69,921 datasets (+36) Structured data, getting broader every week.

Baselight as infrastructure 🦾 corrslight.xyz

Explore the platform and start building: baselight.app

What do 469 billion rows of data look like? They look like agents tracking the pulse of the world in prediction markets and reacting to on-chain transactions settling in near real time. They look like systems analyzing soccer and other sports - every pass, every goal, every transfer. They look like models reading SEC filings - every earnings report, every insider trade. They look like applications reasoning over decades of Federal Reserve indicators. They look like workflows connecting NIH grants to the science of tomorrow. All of it live in one platform. And this is only a subset of what was updated in the last few hours. 70,000+ datasets. 455,000+ tables. From crypto blocks to cancer research. From NFL play-by-play to EU housing prices. This is Baselight - the data platform where agents, analysts, researchers, and builders query the world.

Explore the Global Risk Dashboard here: baselight.app/u/pjsousa/dash… This is just one example - the same data can power your own dashboards, copilots, or agents. Curious to see what you build on top of it 🚀

The same data powering dashboards can power agents. Everything in Baselight isn't just for visualization - it's available to build intelligence layers anywhere. And we keep adding more of it every week: - 465,187,218,680 rows (+7B this week) - 455,190 tables (+213) - 69,885 datasets (+30) That growing foundation is already being used. A new Global Risk Dashboard was built on Baselight - combining inflation, credit markets, energy prices, and prediction markets into a single view. This - and other dashboards - are just examples of what can be built, analyzed, and acted on by agents in near real-time. That’s the direction: continuously expanding data, enabling intelligence everywhere.

Hot take: raw race results don’t tell the full story. The real edge is in: - consistency - pit strategy - marginal gains over time That’s exactly what this dataset unlocks. Explore it yourself: baselight.app

The Ultimate Formula 1 Dataset just dropped on Baselight. We’re adding Formula 1 to our growing sports coverage - alongside soccer, basketball, and NFL. From pole positions to pit stop strategy, every millisecond matters in F1 - and now you can analyze it all in one place. What you get: - Race results & qualifying performance - Driver and constructor standings - Fastest laps & pit stop insights - Circuits, teams, and historical context Whether you're tracking dominance cycles, analyzing race pace, or exploring team strategy - this dataset gives you the full grid.

Financial markets and healthcare data - both expanded this week. Baselight Weekly Update - 458,767,464,260 rows (+7B) - 454,987 tables (+360) - 69,855 datasets (+161) We added two major sources to Baselight: - Hyperliquid: this is a crypto trading platform where people trade assets like Bitcoin, Brent crude oil, and others - similar to a stock exchange, but built on blockchain. Instead of buying the asset itself, users trade perpetual futures - bets on whether the price will go up or down. - HHS health data: this complements other health datasets already on Baselight - including CDC, County Health Rankings, and Eurostat - making it easier to explore public health data in one place. Explore it yourself: - Which asset is being traded the most? - Which health trends stand out in your region?

If you’re working on politics, economics, or investigative journalism - this is a goldmine. Trace donations, uncover patterns, and connect funding flows across decades. Ask your own questions with Baselight AI: baselight.app

+23 billion rows this week - led by the addition of Federal Election Commission (FEC) data. Follow the money. All of it. Baselight Weekly Update (last 7 days) • 451,979,054,166 rows (+23B) • 454,627 tables (+1K) • 69,696 datasets (+283) FEC datasets now live - and the scale is massive • $15.4B in individual contributions (2024) • Up from $14.6B (2020) and ~3× vs 2018 ($5.5B) From individual donors to PACs, electioneering to lobbying: 20 datasets • 352M+ rows • 1980 -> today Who funds American elections? Explore it on Baselight.