--- title: "World V2.0 is now Generally Available: A Point-in-Time Database of Global Events" description: "Explore NOSIBLE World V2.0, a point-in-time global event database with entity sentiment, ticker mapping and classifications for AI and quantitative research." url: "https://nosible.com/blog/world-v2-0-generally-available" --- [NOSIBLE World](https://nosible.com/blog/tag/nosible-world) [Information Retrieval](https://nosible.com/blog/tag/information-retrieval) [Web Search](https://nosible.com/blog/tag/web-search) # World V2.0 is now Generally Available: A Point-in-Time Database of Global Events [Stuart Reid](https://www.linkedin.com/in/stuartgordonreid/) 2026-10-28 8 min read Copy as Markdown ![Diagram showing World V1.2 forking into World V2.0 and its nine processing layers.](https://nosible.com/images/2026/10/nosible-world-v2-fork-hero-icons.png) We envision a world where AI agents connected to the NOSIBLE intelligence stack can see everything happening in the world in real time, as well as everything that has happened throughout history. Currently the stack consists of four levels: 1. Firehose is a near-real-time data feed of what we crawl. 2. Search is a search engine for dated web documents. 3. World is a point-in-time database of worldwide events. 4. Atlas is a point-in-time knowledge graph built from World. Today we're happy to announce that World V2.0 has exited preview and is generally available. Practically speaking this means that the methodology is frozen (to prevent data drift) and V2.0 trial data is available via S3, API, and MCP. In this post, we'll explain what changed from V1.2 to V2.0 and answer common questions. Table of contents [Why We Version the Data](https://nosible.com/blog/world-v2-0-generally-available#why-we-version-the-data) [How Long Will V1.2 Be Maintained?](https://nosible.com/blog/world-v2-0-generally-available#how-long-will-v12-be-maintained) [What's New in World V2.0?](https://nosible.com/blog/world-v2-0-generally-available#whats-new-in-world-v20) [Moving From World V1.2 to V2.0](https://nosible.com/blog/world-v2-0-generally-available#moving-from-world-v12-to-v20) [How does this affect Atlas?](https://nosible.com/blog/world-v2-0-generally-available#how-does-this-affect-atlas) [When Will V3.0 Be Released?](https://nosible.com/blog/world-v2-0-generally-available#when-will-v30-be-released) [Start a World V2.0 Trial](https://nosible.com/blog/world-v2-0-generally-available#start-a-world-v20-trial) ## Why We Version the Data This might sound counterintuitive, but quants prefer consistency over precision. A field that is consistently wrong can still produce a useful signal if its errors are stable. Fix that field halfway through the dataset, however, and a model might interpret the change in the data as a change in the world. That's data drift. That's why we version the entire World dataset and freeze the methodology for each release. Historical and new events are processed consistently, so you can backtest models without changes to our methodology creating artificial shifts in the data. ![Illustration of how changing a data field shifts its distribution and affects a backtest.](https://nosible.com/images/2026/10/nosible-distribution-shift-3640x1560.png) ## How Long Will V1.2 Be Maintained? We will maintain World V1.2 for as long as we maintain V2.0. V2.0 is a fork of V1.2: every V2.0 event starts as a V1.2 event before receiving additional processing to extract relationships. This lets us continue producing V1.2 as part of the same pipeline. You can keep using V1.2 in production and trial V2.0 when you're ready. ## What's New in World V2.0? Select a feature to see how it works. **Named entities** We fine-tuned GLiNER V2.5 to extract and normalize named entities, with types, mention counts, and confidence scores for each entity found in the event text. V1.2 used spaCy to identify named entities, but its results were often imprecise. We needed cleaner names and types for researchers who filter events by entity, and for the models that extract relationships, find locations and map events to tickers. In V2.0, we fine-tuned [GLiNER V2.5](https://huggingface.co/gliner-community/gliner_large-v2.5) to identify people, organizations, products, locations and geopolitical entities, then normalize their names. Each entry in the `entities` field includes a type, mention count and confidence score for the extracted entity. JSON / entities Copy JSON ``` { "entities": { "ORG": [ { "name": "Microsoft", "mention_count": 5, "confidence": 0.9998 }, { "name": "Activision Blizzard", "mention_count": 4, "confidence": 0.9996 } ], "PRODUCT": [ { "name": "Overwatch", "mention_count": 2, "confidence": 0.9883 }, { "name": "Call of Duty", "mention_count": 2, "confidence": 0.9863 }, { "name": "World of Warcraft", "mention_count": 2, "confidence": 0.9851 }, { "name": "Xbox", "mention_count": 3, "confidence": null } ] } } ``` 38 lines **Ticker mapping** We combine company names, ticker symbols, and Atlas relationships with personalized PageRank and reranking to score and explain company matches. An article can be relevant to a company without mentioning its name. News about Zepbound, for example, can matter to Eli Lilly. V1.2 missed some of these connections, leaving researchers without events they could use in company-level signals and backtests. V2.0 matches company names and ticker symbols, then uses Atlas relationships, personalized PageRank and a reranker to find and score additional companies. Each match includes identifiers, a score and an explanation. Read our [ticker mapping analysis](https://nosible.com/blog/the-hidden-path-to-every-ticker). JSON / tickers Copy JSON ``` { "tickers": [ { "ticker": "ATVI.US", "companyv5_id": "c5-8736d2cbf1fc", "name": "Activision Blizzard Inc", "website": "https://activisionblizzard.com", "wiki_qid": "Q344341", "isin": "US00507V1098", "lei": "549300TPQQDBP9GNOF40", "figi": "BBG000CVWGS6", "sector": "Communication Services", "industry": "Entertainment", "country_iso": "US", "exchange": "NASDAQ", "mic": "XNAS", "score": 0.999429, "why": [ { "entity": "Activision Blizzard", "score": 1, "chain": [] }, { "entity": "Call of Duty", "score": 0.489415, "chain": [ { "subject": "Call of Duty", "predicate": "brand_of", "object": "Activision Blizzard", "known_as_of": "2010-03-30" } ] }, { "entity": "Call of Duty Warzone", "score": 0.467143, "chain": [ { "subject": "Activision", "predicate": "develops_product", "object": "Call of Duty Warzone", "known_as_of": "2020-02-17" } ] }, { "entity": "Treyarch", "score": 0.418055, "chain": [ { "subject": "Activision", "predicate": "business_partner_of", "object": "Treyarch", "known_as_of": "2010-04-30" } ] }, { "entity": "Infinity Ward", "score": 0.401962, "chain": [ { "subject": "Activision Blizzard", "predicate": "announced_product", "object": "Call of Duty: Modern Warfare 2", "known_as_of": "2010-01-13" }, { "subject": "Infinity Ward", "predicate": "created", "object": "Call of Duty: Modern Warfare 2", "known_as_of": "2010-01-13" } ] }, { "entity": "World of Warcraft", "score": 0.398586, "chain": [ { "subject": "Activision Blizzard", "predicate": "produces_product", "object": "World of Warcraft", "known_as_of": "2013-05-08" } ] }, { "entity": "Sledgehammer Games", "score": 0.397623, "chain": [ { "subject": "Activision", "predicate": "owns", "object": "Call of Duty: Modern Warfare 3", "known_as_of": "2010-04-10" }, { "subject": "Sledgehammer Games", "predicate": "develops_product", "object": "Call of Duty: Modern Warfare 3", "known_as_of": "2011-01-21" } ] }, { "entity": "Warzone", "score": 0.347064, "chain": [ { "subject": "Activision", "predicate": "develops_product", "object": "Warzone", "known_as_of": "2020-06-02" } ] }, { "entity": "Raven Software", "score": 0.308315, "chain": [ { "subject": "Raven Software", "predicate": "subsidiary_of", "object": "Activision Blizzard", "known_as_of": "2021-12-04" } ] }, { "entity": "Toys for Bob", "score": 0.304981, "chain": [ { "subject": "Activision", "predicate": "announced_product", "object": "Skylanders Trap Team", "known_as_of": "2014-04-23" }, { "subject": "Toys for Bob", "predicate": "announced_product", "object": "Skylanders Trap Team", "known_as_of": "2014-04-23" } ] }, { "entity": "Robert Kotick", "score": 0.258019, "chain": [ { "subject": "Robert Kotick", "predicate": "ceo_of", "object": "Activision Blizzard", "known_as_of": "2010-01-13" } ] }, { "entity": "Beenox", "score": 0.245979, "chain": [ { "subject": "Activision", "predicate": "has_controlling_interest_in", "object": "Beenox", "known_as_of": "2011-01-13" } ] }, { "entity": "High Moon Studios", "score": 0.170555, "chain": [ { "subject": "Activision Publishing, Inc.", "predicate": "collaborates_with", "object": "High Moon Studios", "known_as_of": "2011-02-14" } ] }, { "entity": "Candy Crush", "score": 0.143482, "chain": [ { "subject": "Activision Blizzard", "predicate": "agreed_to_acquire", "object": "King Digital Entertainment", "known_as_of": "2015-11-03" }, { "subject": "King Digital Entertainment", "predicate": "created", "object": "Candy Crush", "known_as_of": "2014-03-28" } ] }, { "entity": "Hearthstone", "score": 0.131669, "chain": [ { "subject": "Activision Blizzard", "predicate": "owns", "object": "Hearthstone", "known_as_of": "2014-11-04" } ] }, { "entity": "Major League Gaming", "score": 0.129653, "chain": [ { "subject": "Activision", "predicate": "business_partner_of", "object": "Major League Gaming", "known_as_of": "2013-02-01" } ] }, { "entity": "Radical Entertainment", "score": 0.123369, "chain": [ { "subject": "Activision", "predicate": "has_controlling_interest_in", "object": "Radical Entertainment", "known_as_of": "2012-06-28" } ] }, { "entity": "Sledgehammer", "score": 0.120041, "chain": [ { "subject": "Sledgehammer", "predicate": "division_of", "object": "Activision", "known_as_of": "2014-02-10" } ] }, { "entity": "No Doubt", "score": 0.060246, "chain": [ { "subject": "Activision Blizzard", "predicate": "reaches_agreement_with", "object": "No Doubt", "known_as_of": "2011-02-16" } ] } ] }, { "ticker": "MSFT.US", "companyv5_id": "c5-37834c8f281e", "name": "Microsoft Corporation", "website": "https://microsoft.com", "wiki_qid": "Q2283", "isin": "US5949181045", "lei": "INR2EJN1ERAN0W5ZP974", "figi": "BBG000BPH459", "sector": "Information Technology", "industry": "Software", "country_iso": "US", "exchange": "NASDAQ", "mic": "XNAS", "score": 0.997681, "why": [ { "entity": "Microsoft", "score": 0.998778, "chain": [] }, { "entity": "Xbox Game Pass", "score": 0.348557, "chain": [ { "subject": "Microsoft", "predicate": "announced_product", "object": "Xbox Game Pass", "known_as_of": "2017-02-28" } ] }, { "entity": "HoloLens", "score": 0.343099, "chain": [ { "subject": "Microsoft", "predicate": "produces_product", "object": "HoloLens", "known_as_of": "2015-01-21" } ] }, { "entity": "Phil Spencer", "score": 0.339706, "chain": [ { "subject": "Phil Spencer", "predicate": "employed_by", "object": "Microsoft", "known_as_of": "2012-03-13" } ] }, { "entity": "Xbox", "score": 0.198376, "chain": [ { "subject": "Microsoft", "predicate": "owns", "object": "Xbox", "known_as_of": "2010-11-09" } ] }, { "entity": "Xbox", "score": 0.176961, "chain": [ { "subject": "Xbox", "predicate": "brand_of", "object": "Microsoft", "known_as_of": "2013-04-18" } ] }, { "entity": "ZeniMax Media", "score": 0.109822, "chain": [ { "subject": "Microsoft", "predicate": "has_controlling_interest_in", "object": "ZeniMax Media", "known_as_of": "2020-09-21" } ] }, { "entity": "Game Pass", "score": 0.050655, "chain": [ { "subject": "Microsoft", "predicate": "operator_of", "object": "Game Pass", "known_as_of": "2019-04-05" } ] }, { "entity": "Satya Narayana Nadella", "score": 0.040199, "chain": [ { "subject": "Satya Narayana Nadella", "predicate": "ceo_of", "object": "Microsoft", "known_as_of": "2019-10-17" } ] }, { "entity": "Microsoft Gaming", "score": 0.029201, "chain": [ { "subject": "Microsoft Gaming", "predicate": "division_of", "object": "Microsoft", "known_as_of": "2021-06-10" } ] }, { "entity": "ZeniMax Media Inc.", "score": 0.022493, "chain": [ { "subject": "Microsoft", "predicate": "agreed_to_acquire", "object": "ZeniMax Media Inc.", "known_as_of": "2020-09-21" } ] } ] }, { "ticker": "6758.TSE", "companyv5_id": "c5-079832f1ee79", "name": "Sony Group Corporation", "website": "https://sony.com", "wiki_qid": "Q41187", "isin": "JP3435000009", "lei": "529900R5WX9N2OI2N910", "figi": "BBG000KBRKP1", "sector": "Consumer Discretionary", "industry": "Household Durables", "country_iso": null, "exchange": "PINK", "mic": null, "score": 0.843019, "why": [ { "entity": "Sony", "score": 0.678525, "chain": [] }, { "entity": "PlayStation 5", "score": 0.476758, "chain": [ { "subject": "Sony", "predicate": "develops_product", "object": "PlayStation 5", "known_as_of": "2016-08-22" } ] }, { "entity": "PlayStation", "score": 0.472379, "chain": [ { "subject": "Sony", "predicate": "produces_product", "object": "PlayStation 3", "known_as_of": "2010-03-01" }, { "subject": "PlayStation", "predicate": "alias_of", "object": "PlayStation 3", "known_as_of": "2010-09-01" } ] }, { "entity": "Sony PS5", "score": 0.089136, "chain": [ { "subject": "Sony", "predicate": "develops_product", "object": "Sony PS5", "known_as_of": "2020-06-15" } ] }, { "entity": "Playstation", "score": 0.0159, "chain": [ { "subject": "Sony", "predicate": "owns", "object": "Sony Playstation", "known_as_of": "2011-11-15" }, { "subject": "Playstation", "predicate": "alias_of", "object": "Sony Playstation", "known_as_of": "2011-11-15" } ] } ] }, { "ticker": "TTWO.US", "companyv5_id": "c5-c0d53969b4c8", "name": "Take-Two Interactive Software Inc", "website": "https://take2games.com", "wiki_qid": "Q94933", "isin": "US8740541094", "lei": "BOMSTHHJK882EWYX3334", "figi": "BBG000BS1YV5", "sector": "Communication Services", "industry": "Entertainment", "country_iso": "US", "exchange": "NASDAQ", "mic": "XNAS", "score": 0.833647, "why": [ { "entity": "Take-Two Interactive", "score": 0.981436, "chain": [] }, { "entity": "Zynga", "score": 0.612372, "chain": [ { "subject": "Take-Two Interactive Software Inc.", "predicate": "completed_acquisition_of", "object": "Zynga", "known_as_of": "2022-01-10" } ] } ] }, { "ticker": "0700.HK", "companyv5_id": "c5-f368f269812f", "name": "Tencent Holdings Limited", "website": "https://tencent.com", "wiki_qid": "Q860580", "isin": "KYG875721634", "lei": "254900N4SLUMW4XUYY11", "figi": "BBG001S8C4T6", "sector": null, "industry": null, "country_iso": null, "exchange": "HK", "mic": null, "score": 0.498838, "why": [ { "entity": "Tencent", "score": 0.922048, "chain": [] } ] } ] } ``` 551 lines **Event classifications** We combine EmbeddingGemma with trained classification heads to assign ranked event categories and probabilities across all 15 taxonomies in World. V1.2 classified events using embedding similarity, which produced noisy ontology labels. Researchers need more precise classifications to select events by topic or industry, set confidence thresholds and assemble datasets for investment research. V2.0 uses [EmbeddingGemma](https://developers.googleblog.com/en/introducing-embeddinggemma/) and trained heads across 15 taxonomies. It returns categories and probabilities in JSON. In V2.0 we keep only classifications above a 50% confidence to reduce JSON payload size. JSON / ontologies Copy JSON ``` { "ontologies": [ { "ontology": "iab", "tier_1": "Business and Finance", "tier_2": "Business", "tier_3": "Business Banking & Finance", "tier_4": "Mergers and Acquisitions", "probability": 0.9926 }, { "ontology": "schema_org_event", "label": "BusinessEvent", "probability": 0.9733 }, { "ontology": "gics", "probability": 0.9585, "sector": "Communication Services", "industry_group": "Media & Entertainment", "industry": "Entertainment", "sub_industry": "Interactive Home Entertainment" }, { "ontology": "media_frame", "label": "Economic Consequences", "probability": 0.9333 }, { "ontology": "nosible", "category": "Capital Markets", "event": "Acquisition Concluded", "subcategory": "Acquistion", "probability": 0.9205 }, { "ontology": "asset_class", "main_class": "Equities", "sub_class": "Public Equities", "probability": 0.757 }, { "ontology": "ekman7_emotion", "label": "fear", "probability": 0.6716 } ] } ``` 48 lines **Entity sentiments** We use point-in-time ChronoGPT vintages to score whether an event is good or bad for each entity, with entity-specific baseline corrections and historical normalization. One event can affect the entities involved differently. In a lawsuit, for example, a ruling may be good for one company and bad for another. An event-level sentiment score alone cannot express both assessments for researchers studying the parties involved. We use a [ChronoGPT](https://huggingface.co/collections/manelalab/chronogpt) vintage trained on information available before each event to score whether it is good or bad for each named entity. We correct for an entity-specific baseline and normalize using historical data available before the event. This makes the scores point-in-time safe. The `sentiments` field gives each entity a label and standardized score for the event. JSON / sentiments Copy JSON ``` { "sentiments": { "ORG": [ { "name": "Microsoft", "sentiment_label": "positive", "sentiment_score": 1.2614899910817696 }, { "name": "Activision Blizzard", "sentiment_label": "neutral", "sentiment_score": -0.17186529086908467 } ], "PRODUCT": [ { "name": "Overwatch", "sentiment_label": "neutral", "sentiment_score": -0.18561534589949163 }, { "name": "Call of Duty", "sentiment_label": "neutral", "sentiment_score": -0.26792553853293627 }, { "name": "World of Warcraft", "sentiment_label": "neutral", "sentiment_score": -0.1416127244272873 }, { "name": "Xbox", "sentiment_label": "neutral", "sentiment_score": -0.0014365055764485465 } ] } } ``` 38 lines **Entity relationships** We fine-tuned GLiNER to extract entity relationships, then rerank each triple; Atlas aggregates those relationships into a point-in-time knowledge graph. Names alone don't tell you how entities are connected. Researchers and AI agents also need to know, for example, which company owns a product or has agreed to acquire another company. Those relationships also build the Atlas knowledge graph. We fine-tuned [GLiNER](https://www.knowledgator.com/research/gliner-relex) to extract relationships and built an internal reranker to assess them. World stores each result as a subject–predicate–object triple with scores from both models. Atlas aggregates relationships across events into a [point-in-time knowledge graph](https://nosible.com/blog/announcing-atlas-the-point-in-time-knowledge-graph). JSON / triples Copy JSON ``` { "triples": [ { "subject": "Microsoft", "subject_type": "ORG", "predicate": "agreed_to_acquire", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.99951171875, "reranker_score": 0.999853491783142 }, { "subject": "Call of Duty", "subject_type": "PRODUCT", "predicate": "brand_of", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.8369140625, "reranker_score": 0.9959771037101746 }, { "subject": "Microsoft", "subject_type": "ORG", "predicate": "has_controlling_interest_in", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.978515625, "reranker_score": 0.9943031072616576 }, { "subject": "Activision Blizzard", "subject_type": "ORG", "predicate": "reaches_agreement_with", "object": "Microsoft", "object_type": "ORG", "original_score": 0.9951171875, "reranker_score": 0.9934230446815492 }, { "subject": "Activision Blizzard", "subject_type": "ORG", "predicate": "signed_contract_with", "object": "Microsoft", "object_type": "ORG", "original_score": 0.98681640625, "reranker_score": 0.988917112350464 }, { "subject": "Microsoft", "subject_type": "ORG", "predicate": "proposed_acquisition_of", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.9716796875, "reranker_score": 0.9835337996482848 }, { "subject": "Activision Blizzard", "subject_type": "ORG", "predicate": "produces_product", "object": "Call of Duty", "object_type": "PRODUCT", "original_score": 0.73193359375, "reranker_score": 0.9816126823425292 }, { "subject": "Overwatch", "subject_type": "PRODUCT", "predicate": "brand_of", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.85205078125, "reranker_score": 0.96987122297287 }, { "subject": "World of Warcraft", "subject_type": "PRODUCT", "predicate": "brand_of", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.8681640625, "reranker_score": 0.9692660570144652 }, { "subject": "Microsoft", "subject_type": "ORG", "predicate": "owns_equity_stake_in", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.9638671875, "reranker_score": 0.9489645957946776 }, { "subject": "Microsoft", "subject_type": "ORG", "predicate": "owns", "object": "Overwatch", "object_type": "PRODUCT", "original_score": 0.6181640625, "reranker_score": 0.9377830028533936 }, { "subject": "Microsoft", "subject_type": "ORG", "predicate": "planned_investment_in", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.86376953125, "reranker_score": 0.9242786765098572 }, { "subject": "Activision Blizzard", "subject_type": "ORG", "predicate": "produces_product", "object": "World of Warcraft", "object_type": "PRODUCT", "original_score": 0.6640625, "reranker_score": 0.9060314297676086 }, { "subject": "Microsoft", "subject_type": "ORG", "predicate": "owns", "object": "Call of Duty", "object_type": "PRODUCT", "original_score": 0.6552734375, "reranker_score": 0.89674973487854 }, { "subject": "Activision Blizzard", "subject_type": "ORG", "predicate": "negotiates_with", "object": "Microsoft", "object_type": "ORG", "original_score": 0.9345703125, "reranker_score": 0.8603691458702087 }, { "subject": "Activision Blizzard", "subject_type": "ORG", "predicate": "produces_product", "object": "Overwatch", "object_type": "PRODUCT", "original_score": 0.63037109375, "reranker_score": 0.859545886516571 }, { "subject": "Microsoft", "subject_type": "ORG", "predicate": "completed_acquisition_of", "object": "Activision Blizzard", "object_type": "ORG", "original_score": 0.76025390625, "reranker_score": 0.8193880319595337 }, { "subject": "Activision Blizzard", "subject_type": "ORG", "predicate": "publishes", "object": "Call of Duty", "object_type": "PRODUCT", "original_score": 0.7646484375, "reranker_score": 0.5980334281921387 } ] } ``` 166 lines **English translations** We use TranslateGemma to translate event titles, descriptions, and key points into English while retaining the original text alongside them to check translations. World contains events reported in many languages. Researchers working in English need a way to analyze those events together, search their contents and compare findings across markets while retaining access to the original text for inspection. We use Google's [TranslateGemma](https://blog.google/innovation-and-ai/technology/developers-tools/translategemma/) to translate non-English event titles, descriptions and key points into English. The original text remains in `event`, while the translation is stored in `english`. For events originally in English, the field is unchanged from the source text. JSON / english Copy JSON ``` { "english": { "date": "2022-01-18", "language": "en", "title": "Microsoft Acquires Activision Blizzard in $68.7 Billion Gaming Deal", "description": "Microsoft has agreed to acquire Activision Blizzard for approximately $68.7 billion, marking a massive consolidation in the global gaming industry. This transaction brings iconic franchises like Call of Duty, World of Warcraft, and Overwatch under Microsoft's Xbox division. The deal aims to expand the Game Pass library while ensuring Activision titles remain available across multiple platforms. Regulatory scrutiny and worker concerns remain key factors as the acquisition proceeds.", "points": [ "Microsoft acquired Activision Blizzard for approximately $68.7 billion.", "The acquisition marks a massive consolidation in the global gaming industry.", "The deal brings Call of Duty, World of Warcraft, and Overwatch under Microsoft's Xbox division.", "The transaction aims to expand the Xbox Game Pass library.", "Activision Blizzard has been engulfed in controversy.", "Regulatory scrutiny and worker concerns remain key factors in the acquisition." ] } } ``` 16 lines **Multiple locations** We identify an event’s primary location and resolve named places and geopolitical entities to coordinates, recording how each was found in the original event text. Events can involve more than one place. A single location can miss the geographic scope of a cross-border development and leave researchers unable to find an event when searching for another country or geographic region mentioned in the event. We identify an event's primary geographic context from its title and description, then resolve named locations and geopolitical entities to coordinates. The `locations` field records each place, how it was found and which is the primary location for the event. JSON / locations Copy JSON ``` { "locations": [ { "continent": "North America", "region": "North America", "country": "United States", "lat": 37.422, "lng": -122.0841, "source": "llm", "matched_on": "{event.title}+{event.description}" } ] } ``` 13 lines **Event key points** We extract key points from event sources, capturing named entities and relationships alongside each event's title and description, including the companies and products involved. An event's title and description don't always capture every detail a researcher or AI agent needs. Key points add facts about the people, companies and products involved, including how they are connected. We added `event.points`, a list of factual points drawn from an event's sources. The points sit alongside the title and description in the original `event` field, and their English translations are available in `english` for analysis across languages and markets. JSON / event.points Copy JSON ``` { "event": { "points": [ "Microsoft acquired Activision Blizzard for approximately $68.7 billion.", "The acquisition marks a massive consolidation in the global gaming industry.", "The deal brings Call of Duty, World of Warcraft, and Overwatch under Microsoft's Xbox division.", "The transaction aims to expand the Xbox Game Pass library.", "Activision Blizzard has been engulfed in controversy.", "Regulatory scrutiny and worker concerns remain key factors in the acquisition." ] } } ``` 12 lines **Availability** We record hourly source-document counts in a top-level field and remove raw timestamps that slowed processing of large events with extensive source coverage. Large events can be reported by thousands of sources. In V1.2, retaining raw timestamps could make event records large and slow to process, while researchers still needed to know when the coverage appeared and how it developed over the lifetime of the event. V2.0 removes the raw `timestamps` field and adds `availability` as a top-level field. It records source-document counts by hour, giving researchers a time series of coverage that they can use without carrying and processing every individual timestamp. JSON / availability Copy JSON ``` { "availability": { "2022-01-18T04:00:00+00:00": 1, "2022-01-18T05:00:00+00:00": 1, "2022-01-18T06:00:00+00:00": 2, "2022-01-18T07:00:00+00:00": 1, "2022-01-18T08:00:00+00:00": 6, "2022-01-18T09:00:00+00:00": 6, "2022-01-18T10:00:00+00:00": 25, "2022-01-18T11:00:00+00:00": 8, "2022-01-18T12:00:00+00:00": 4, "2022-01-18T13:00:00+00:00": 162, "2022-01-18T14:00:00+00:00": 170, "2022-01-18T15:00:00+00:00": 100, "2022-01-18T16:00:00+00:00": 73, "2022-01-18T17:00:00+00:00": 75, "2022-01-18T18:00:00+00:00": 50, "2022-01-18T19:00:00+00:00": 45, "2022-01-18T20:00:00+00:00": 47, "2022-01-18T21:00:00+00:00": 37, "2022-01-18T22:00:00+00:00": 29, "2022-01-18T23:00:00+00:00": 39, "2022-01-19T00:00:00+00:00": 1, "2022-01-19T01:00:00+00:00": 2, "2022-01-19T02:00:00+00:00": 1, "2022-01-19T03:00:00+00:00": 1, "2022-01-19T13:00:00+00:00": 1, "2022-01-18T00:00:00+00:00": 0, "2022-01-18T01:00:00+00:00": 0, "2022-01-18T02:00:00+00:00": 0, "2022-01-18T03:00:00+00:00": 0 } } ``` 33 lines ## Moving From World V1.2 to V2.0 We made V2.0 a fork of V1.2 to make migration as straightforward as possible. V2.0 retains many event fields, but entity extraction, ontology classifications and ticker mappings have changed. The comparison summarizes the changes between V1.2 and V2.0. Before updating your pipeline, compare the fields and companies you use, and check the production data dictionary and a sample of your delivery. | Field | V1.2 | V2.0 | What changed | | --- | --- | --- | --- | | `english` | — | Added | English translations of titles, descriptions and points. | | `availability` | — | Added | Hourly source counts promoted from `timestamps`. | | `locations` | — | Added | Multiple geographic locations for each event. | | `sentiments` | — | Added | Sentiment for each named entity in the event. | | `triples` | — | Added | Relationships between named entities in the event. | | `timestamps` | Present | Removed | Hourly counts retained in `availability`. | | `coordinate` | Present | Removed | Coordinates now stored in `locations`. | | `extra` | Present | Removed | Signal probabilities moved to `signals`. | | `entities` | Present | Updated | GLiNER replaces spaCy; JSON representation changes. | | `ontology` / `ontologies` | Present | Updated | Renamed; classification and JSON change. | | `event` | Present | Expanded | Adds `points` containing key event facts. | | `coverage` | Present | Expanded | Adds first, last and most-seen timestamps. | | `signals` | Present | Expanded | Adds sentiment and forward-looking probabilities. | | `tickers` | Present | Expanded | Adds match scores and explanations for each company. | | `event_id` | Present | Unchanged | No changes were made from V1.2 to V2.0. | | `version` | Present | Unchanged | No changes were made from V1.2 to V2.0. | | `has_tickers` | Present | Unchanged | No changes were made from V1.2 to V2.0. | | `provenance` | Present | Unchanged | No changes were made from V1.2 to V2.0. | | `oai_vector` | Present | Unchanged | No changes were made from V1.2 to V2.0. | ## How does this affect Atlas? World describes events as they happen. Atlas connects the entities and relationships found in those events into a point-in-time knowledge graph. Each World V2.0 event has a `triples` field containing relationships extracted from the event text. A triple has three parts: a subject, a predicate and an object. For example, `Sprite → brand_of → Coca-Cola` says that Sprite is a brand of Coca-Cola. Each triple includes entity types, extraction scores and reranker scores. Every day, we aggregate triples across World events, resolve names to normalized entities and use independent sources to corroborate relationships. Those relationships become edges in Atlas. The graph records when relationships became known and retains their evidence, so you can reconstruct what was known on any date. ## When Will V3.0 Be Released? We plan to release World V3.0 in Q1 or Q2 of 2027. The goal is to make the dataset bigger by bringing in more sources and more of the web. ## Start a World V2.0 Trial If you're using World, start with events and companies you know. Compare V1.2 and V2.0 ticker mappings, entity extraction and ontology classifications. Test entity-level sentiment for companies across your investment universe. I'm excited about the research these additions make possible, especially using entity-level sentiment to study how the same events affect different companies. [Start a World V2.0 trial](https://nosible.com/start-trial) or talk to us about testing it with your investment universe. [All Research](https://nosible.com/blog) Related Research ![Dotted NOSIBLE map showing seven oil chokepoint polygons and 5,261 matching World events for September 2026.](https://nosible.com/images/2026/10/world-geography-chokepoints.png) [NOSIBLE World](https://nosible.com/blog/tag/nosible-world) [Information Retrieval](https://nosible.com/blog/tag/information-retrieval) [Web Search](https://nosible.com/blog/tag/web-search) ### [Polygon Queries: Intelligence Across Space and Time](https://nosible.com/blog/polygon-queries-geospatial-temporal-intelligence) 2026-10-16 8 min read ![The NOSIBLE Relevance Lab showing what Atlas links to Eli Lilly, with selpercatinib traced to Eli Lilly through Loxo Oncology.](https://nosible.com/images/2026/10/atlas-tickers-relevance-lab-2000x857.png) [NOSIBLE World](https://nosible.com/blog/tag/nosible-world) [Information Retrieval](https://nosible.com/blog/tag/information-retrieval) [Web Search](https://nosible.com/blog/tag/web-search) [Trading Signals](https://nosible.com/blog/tag/trading-signals) ### [The Hidden Path to Every Ticker: 15% to 22% Coverage, Zero Look-Ahead](https://nosible.com/blog/the-hidden-path-to-every-ticker) 2026-10-06 11 min read ![Compression is intelligence: NOSIBLE Search indexes the web, World produces events, and Atlas maps facts into a knowledge graph.](https://nosible.com/images/2026/09/atlas-nosible-stack-3640x1560.png) [NOSIBLE World](https://nosible.com/blog/tag/nosible-world) [Information Retrieval](https://nosible.com/blog/tag/information-retrieval) [Web Search](https://nosible.com/blog/tag/web-search) ### [Announcing Atlas: The Point-in-time Knowledge Graph](https://nosible.com/blog/announcing-atlas-the-point-in-time-knowledge-graph) 2026-09-30 20 min read > Explore NOSIBLE World V2.0, a point-in-time global event database with entity sentiment, ticker mapping and classifications for AI and quantitative research. **URL:** https://nosible.com/blog/world-v2-0-generally-available