Cryptocurrency historical datasets from January 2012 (if available) to October 2021 were obtained and integrated from various sources and Application Programming Interfaces (APIs) including Yahoo Finance, Cryptodownload, CoinMarketCap, various Kaggle datasets, and multiple APIs. While these datasets used various formats of time (e.g., minutes, hours, days), in order to integrate the datasets days format was used for in this research study. The integrated cryptocurrency historical datasets for 80 cryptocurrencies including but not limited to Bitcoin (BTC), Ethereum (ETH), Binance Coin (BNB), Cardano (ADA), Tether (USDT), Ripple (XRP), Solana (SOL), Polkadot (DOT), USD Coin (USDC), Dogecoin (DOGE), Tron (TRX), Bitcoin Cash (BCH), Litecoin (LTC), EOS (EOS), Cosmos (ATOM), Stellar (XLM), Wrapped Bitcoin (WBTC), Uniswap (UNI), Terra (LUNA), SHIBA INU (SHIB), and 60 more cryptocurrencies were uploaded in this online Mendeley data repository. Although the primary attribute of including the mentioned cryptocurrencies was the Market Capitalization, a subject matter expert i.e., a professional trader has also guided the initial selection of the cryptocurrencies by analyzing various indicators such as Relative Strength Index (RSI), Moving Average Convergence/Divergence (MACD), MYC Signals, Bollinger Bands, Fibonacci Retracement, Stochastic Oscillator and Ichimoku Cloud. The primary features of this dataset that were used as the decision-making criteria of the CLUS-MCDA II approach are Timestamps, Open, High, Low, Closed, Volume (Currency), % Change (7 days and 24 hours), Market Cap and Weighted Price values. The available excel and CSV files in this data set are just part of the integrated data and other databases, datasets and API References that was used in this study are as follows: [1] https://finance.yahoo.com/ [2] https://coinmarketcap.com/historical/ [3] https://cryptodatadownload.com/ [4] https://kaggle.com/philmohun/cryptocurrency-financial-data [5] https://kaggle.com/deepshah16/meme-cryptocurrency-historical-data [6] https://kaggle.com/sudalairajkumar/cryptocurrencypricehistory [7] https://min-api.cryptocompare.com/data/price?fsym=BTC&tsyms=USD [8] https://min-api.cryptocompare.com/ [9] https://p.nomics.com/cryptocurrency-bitcoin-api [10] https://www.coinapi.io/ [11] https://www.coingecko.com/en/api [12] https://cryptowat.ch/ [13] https://www.alphavantage.co/ This dataset is part of the CLUS-MCDA (Cluster analysis for improving Multiple Criteria Decision Analysis) and CLUS-MCDAII Project: https://aimaghsoodi.github.io/CLUSMCDA-R-Package/ https://github.com/Aimaghsoodi/CLUS-MCDA-II https://github.com/azadkavian/CLUS-MCDA

Crypto Fear and Greed Index

Each day, the website https://alternative.me/crypto/fear-and-greed-index/ publishes this index based on analysis of emotions and sentiments from different sources crunched into one simple number: The Fear & Greed Index for Bitcoin and other large cryptocurrencies.

Why Measure Fear and Greed?

The crypto market behaviour is very emotional. People tend to get greedy when the market is rising which results in FOMO (Fear of missing out). Also, people often sell their coins in irrational reaction of seeing red numbers. With our Fear and Greed Index, we try to save you from your own emotional overreactions. There are two simple assumptions:

• Extreme fear can be a sign that investors are too worried. That could be a buying opportunity.
• When Investors are getting too greedy, that means the market is due for a correction.

Therefore, we analyze the current sentiment of the Bitcoin market and crunch the numbers into a simple meter from 0 to 100. Zero means "Extreme Fear", while 100 means "Extreme Greed". See below for further information on our data sources.

Data Sources

We are gathering data from the five following sources. Each data point is valued the same as the day before in order to visualize a meaningful progress in sentiment change of the crypto market.

First of all, the current index is for bitcoin only (we offer separate indices for large alt coins soon), because a big part of it is the volatility of the coin price.

But let’s list all the different factors we’re including in the current index:

Volatility (25 %)

We’re measuring the current volatility and max. drawdowns of bitcoin and compare it with the corresponding average values of the last 30 days and 90 days. We argue that an unusual rise in volatility is a sign of a fearful market.

Market Momentum/Volume (25%)

Also, we’re measuring the current volume and market momentum (again in comparison with the last 30/90 day average values) and put those two values together. Generally, when we see high buying volumes in a positive market on a daily basis, we conclude that the market acts overly greedy / too bullish.

Social Media (15%)

While our reddit sentiment analysis is still not in the live index (we’re still experimenting some market-related key words in the text processing algorithm), our twitter analysis is running. There, we gather and count posts on various hashtags for each coin (publicly, we show only those for Bitcoin) and check how fast and how many interactions they receive in certain time frames). A unusual high interaction rate results in a grown public interest in the coin and in our eyes, corresponds to a greedy market behaviour.

Surveys (15%) currently paused

Together with strawpoll.com (disclaimer: we own this site, too), quite a large public polling platform, we’re conducting weekly crypto polls and ask people how they see the market. Usually, we’re seeing 2,000 - 3,000 votes on each poll, so we do get a picture of the sentiment of a group of crypto investors. We don’t give those results too much attention, but it was quite useful in the beginning of our studies. You can see some recent results here.

Dominance (10%)

The dominance of a coin resembles the market cap share of the whole crypto market. Especially for Bitcoin, we think that a rise in Bitcoin dominance is caused by a fear of (and thus a reduction of) too speculative alt-coin investments, since Bitcoin is becoming more and more the safe haven of crypto. On the other side, when Bitcoin dominance shrinks, people are getting more greedy by investing in more risky alt-coins, dreaming of their chance in next big bull run. Anyhow, analyzing the dominance for a coin other than Bitcoin, you could argue the other way round, since more interest in an alt-coin may conclude a bullish/greedy behaviour for that specific coin.

Trends (10%)

We pull Google Trends data for various Bitcoin related search queries and crunch those numbers, especially the change of search volumes as well as recommended other currently popular searches. For example, if you check Google Trends for "Bitcoin", you can’t get much information from the search volume. But currently, you can see that there is currently a +1,550% rise of the query „bitcoin price manipulation“ in the box of related search queries (as of 05/29/2018). This is clearly a sign of fear in the market, and we use that for our index.

There's a story behind every dataset and here's your opportunity to share yours.

Copyright disclaimer

This dataset is produced and maintained by the administrators of https://alternative.me/crypto/fear-and-greed-index/.

This published version is an unofficial copy of their data, which can be also collected using their API (e.g., GET https://api.alternative.me/fng/?limit=10&format=csv&date_format=us).

I am a new developer and I would greatly appreciate your support. If you find this dataset helpful, please consider giving it an upvote!

Key Features:

Complete 1d Data: Raw 1d historical data from multiple exchanges, covering the entire trading history of BTCUSD available through their API endpoints. This dataset is updated daily to ensure up-to-date coverage.

Combined Index Dataset: A unique feature of this dataset is the combined index, which is derived by averaging all other datasets into one, please see attached notebook. This creates the longest continuous, unbroken BTCUSD dataset available on Kaggle, with no gaps and no erroneous values. It gives a much more comprehensive view of the market i.e. total volume across multiple exchanges.

Superior Performance: The combined index dataset has demonstrated superior 'mean average error' (MAE) metric performance when training machine learning models, compared to single-source datasets by a whole order of MAE magnitude.

Unbroken History: The combined dataset's continuous history is a valuable asset for researchers and traders who require accurate and uninterrupted time series data for modeling or back-testing.

https://i.imgur.com/8pw9H5E.png" alt="BTCUSD Dataset Summary">

https://i.imgur.com/AuFSkzb.png" alt="Combined Dataset Close Plot"> This plot illustrates the continuity of the dataset over time, with no gaps in data, making it ideal for time series analysis.

Included Resources:

Two Notebooks:

Dataset Usage and Diagnostics: This notebook demonstrates how to use the dataset and includes a powerful data diagnostics function, which is useful for all time series analyses.

Aggregating Multiple Data Sources: This notebook walks you through the process of combining multiple exchange datasets into a single, clean dataset. (Currently unavailable, will be added shortly)

Overview This dataset provides daily snapshots of cryptocurrency, stock market, and forex data.

Sources Yahoo Finance (via yfinance)

Features Automated daily updates Covers major global indices and top cryptocurrencies Includes sentiment analysis for financial news

Use Cases Financial market analysis Machine learning for price prediction Trading strategy research

License Data compiled from public APIs for educational and analytical use.

The dataset was collected for the period spanning between 01/07/2019 and 31/12/2022.The historical Twitter volume were retrieved using ‘‘Bitcoin’’ (case insensitive) as the keyword from bitinfocharts.com. Google search volume was retrieved using library Gtrends. 2000 tweets per day using 4 times interval were crawled by employing Twitter API with the keyword “Bitcoin. The daily closing prices of Bitcoin, oil price, gold price, and U.S stock market indexes (S&P 500, NASDAQ, and Dow Jones Industrial Average) were collected using R libraries either Quantmod or Quandl.

Two of the most common questions that are often asked about cryptoassets are 'what is the intrinsic value (IV) of a cryptoasset?', and 'what makes one crypto asset more valuable that another asset?'.

In a complex market with literally thousands of instruments, RDA Ranking clarifies the intrinsic value of cryptoassets.

The index uses a proprietary algorithm to analyse asset attributes and compute their intrinsic value on 0 to N point scale - where 0 indicates no intrinsic value and N is the highest intrinsic value for a given asset.

The Market IV Level is defined as the maximum RDA points of all assets at any given point. The Market IV Level serves as a reference for the evolution of fundamental drivers of the cryptoasset industry. By definition, it is the higher frontier of intrinsic value of cryptoassets.

Dataset Overview Each row represents a unique log emitted during transaction execution: • Canonical positioning: (block_number, tx_index, log_index) • Emitting contract address • Primary event topic (topic_zero) • Additional topics (data_topics) • Raw event data payload

All fields are stored exactly as produced by the node, with direct RLP verifiability for topics, data, and contract address.

Every log includes a deterministic _tracing_id that links the record to its genesis event and upstream transaction, forming the foundation for decoded events, swaps, liquidity, NFT events, and custom protocol decoders in downstream BlockDB products.

Chains and Coverage ETH, BSC, Base, Arbitrum, Unichain, Avalanche, Polygon, Celo, Linea, Optimism (others on request). Full history from chain genesis; reorg-aware real-time ingestion and updates.

Schema List of columns exactly as delivered: • block_number BIGINT – Block number that contains the emitting transaction • tx_index INTEGER – Zero-based index of the transaction within the block • log_index INTEGER – Zero-based position of the log within the transaction • contract_address BYTEA – 20-byte address of the contract that emitted the log • topic_zero BYTEA – 32-byte primary topic hash identifying the event type (NULL for anonymous events) • data_topics BYTEA[] – Array of additional topics (topics[1..n]), as raw bytes • data BYTEA – Raw event data payload as emitted on-chain • _tracing_id BYTEA – Deterministic lineage identifier of this log record • _created_at TIMESTAMPTZ – Record creation timestamp • _updated_at TIMESTAMPTZ – Record last update timestamp

Notes • Primary key: (block_number, tx_index, log_index) guarantees canonical ordering and uniqueness. • Foreign key: (block_number, tx_index) links each log directly to its canonical transaction record. • Indexes on contract_address, topic_zero, and (contract_address, topic_zero) support fast protocol- or event-specific scans. • Binary values can be rendered as hex via encode(column, 'hex') in SQL for display or downstream joins.

Lineage & Integrity Direct RLP-verifiable fields: contract_address, topic_zero, data_topics, data, and log_index are all directly or indirectly validated against node RLP.

_tracing_id provides a deterministic, cryptographic handle for each log row, enabling: • Provenance tracking from raw logs to decoded events and higher-level features • Reproducible analytics and signal extraction • Cross-system consistency checks (RPC vs. indexers vs. internal warehouses)

Common Use Cases • Building decoded event layers (swaps, LP actions, mints, burns, governance events, NFT activity) • Reconstructing DEX activity and liquidity flows directly from raw logs • Protocol-specific analytics (AMMs, lending, perpetuals, bridges, staking) from first principles • Detecting MEV patterns, liquidations, and arbitrage events at log-level resolution

Quality • Verifiable lineage: deterministic cryptographic hashes per row • Reorg-aware ingestion: continuity and consistency across forks • Complete historical coverage: from chain genesis to present

This Dataset is described in Charting the Landscape of Online Cryptocurrency Manipulation. IEEE Access (2020), a study that aims to map and assess the extent of cryptocurrency manipulations within and across the online ecosystems of Twitter, Telegram, and Discord. Starting from tweets mentioning cryptocurrencies, we leveraged and followed invite URLs from platform to platform, building the invite-link network, in order to study the invite link diffusion process.

Please, refer to the paper below for more details.

Nizzoli, L., Tardelli, S., Avvenuti, M., Cresci, S., Tesconi, M. & Ferrara, E. (2020). Charting the Landscape of Online Cryptocurrency Manipulation. IEEE Access (2020).

This dataset is composed of:

• ~16M tweet ids shared between March and May 2019, mentioning at least one of the 3,822 cryptocurrencies (cashtags) provided by the CryptoCompare public API;
• ~13k nodes of the invite-link network, i.e., the information about the Telegram/Discord channels and Twitter users involved in the cryptocurrency discussion (e.g., id, name, audience, invite URL);
• ~62k edges of the invite-link network, i.e., the information about the flow of invites (e.g., source id, target id, weight).

With such information, one can easily retrieve the content of channels and messages through Twitter, Telegram, and Discord public APIs.

Please, refer to the README file for more details about the fields.

3MEth Dataset OverviewSection 1: Token TransactionsThis section provides 303 million transaction records from 3,880 tokens and 35 million users on the Ethereum blockchain. The data is stored in 3,880 CSV files, each representing a specific token. Each transaction includes the following information:Sender and receiver wallet addresses: Enables network analysis and user behavior studies.Token address: Links transactions to specific tokens for token-specific analysis.Transaction value: Reflects the number of tokens transferred, essential for liquidity studies.Blockchain timestamp: Captures transaction timing for temporal analysis.Apart from the large dataset, we also provide a smaller CSV file containing 267,242 transaction records from 29,164 wallet addresses. This smaller dataset involves a total of 1,194 tokens, covering the time period September 2016 to November 2023. This detailed transaction data is critical for studying user behavior, liquidity patterns, and tasks such as link prediction and fraud detection.Section 2: Token InformationThis section offers metadata for 3,880 tokens, stored in corresponding CSV files. Each file contains:Timestamp: Marks the time of data update.Token price: Useful for price prediction and volatility studies.Market capitalization: Reflects the token's market size and dominance.24-hour trading volume: Indicates liquidity and trading activity.Section 3: Global Market IndicesThis section provides macro-level data to contextualize token transactions, stored in separate CSV files. Key indicators include:Bitcoin dominance: Tracks Bitcoin's share of the cryptocurrency market.Total market capitalization: Measures the overall market's value, with breakdowns by token type.Stablecoin market capitalization: Highlights stablecoin liquidity and stability.24-hour trading volume: A key measure of market activity.These indices are essential for integrating global market trends into predictive models for volatility and risk-adjusted returns.Section 4: Textual IndicesThis section contains sentiment data from Reddit's Ethereum community, covering 7,800 top posts from 2014 to 2024. Each post includes:Post score (net upvotes): Reflects engagement and sentiment strength.Timestamp: Aligns sentiment with price movements.Number of comments: Gauges sentiment intensity.Sentiment indices: Sentiment scores computed using methods detailed in the data preprocessing section.The full Reddit textual dataset is available upon request; please contact us for access. Alternatively our open-source repository includes a tool to guide users in collecting Reddit data. Researchers are encouraged to apply for a Reddit API Key and adhere to Reddit's policies. This data is valuable for understanding social dynamics in the market and enhancing sentiment analysis models that can explain market movements and improve behavioral predictions.

The attached files enable the full replication of the empirical analysis. All .ipynb files ending in _retrieve contain the code for retrieving data via the respective APIs.

We manually obtained the total global crypto asset market capitalization, as well as the market capitalization of each individual cryptocurrency and stablecoin, from CoinGecko. The data is stored as individual .csv files, which are imported into the .ipynb files. Additionally, we retrieve the following daily variables from DataStream: the TRY/USD exchange rate, MSCI World Index, TR2YT, US2YT, BIST 100, and S&P 500.

The dummy variable indicating month-to-month inflation shocks exceeding 5% is based on the file AppendixB_Türkiye_Inflation_figures_CSV_format.csv. The data in this file was manually obtained from the website of the Central Bank of the Republic of Türkiye (CBRT).

Descriptive statistics and empirical analysis are conducted in Stata. All figures are generated using the .ipynb file P3_Code_Python_figures.

Note that the categorization used in the published paper (MegaCap, LargeCap, MidCap), which follows FINRA’s classification of stocks by market capitalization, differs from the terminology in the code (LargeCap, MidCap, SmallCap). This discrepancy is purely nominal and does not affect the results.

Dataset Overview Canonical registry of all DEX liquidity pools, normalized into the BlockDB schema for deterministic cross-chain analysis. Fees are modeled separately as versioned terms, so fee changes over time are tracked precisely and are joinable back to the pool.

Chains and Coverage ETH, BSC, Base, Arbitrum, Unichain, Avalanche, Polygon, Celo, Linea, Optimism (others on request). Full history from chain genesis; reorg-aware real-time ingestion and updates. Coverage includes: • Uniswap V2, V3, V4 • Balancer V2, PancakeSwap, Solidly, Maverick, Aerodrome, and others

Schema List the columns exactly as delivered.

liquidity_pools (base registry) • uid BIGINT NOT NULL - stable pool identifier (derived from address or pool_id) • block_number BIGINT - first block where the token was recognized • block_time TIMESTAMPTZ - UTC timestamp when the block was mined • tx_index INTEGER - tx index for that event • log_index INTEGER - log index for that event • contract_address BYTEA NULL - 20-byte pool address (v2/v3-style) • pool_id BYTEA NULL - 32-byte pool id (v4-style / manager-based) • factory BYTEA NOT NULL - DEX factory / pool manager address • type_id INTEGER NOT NULL - pool type FK (constant-product, concentrated, stable/weighted, etc.) • pairnum NUMERIC(6) NULL - optional pair ordinal/descriptor • tokens BYTEA[] NOT NULL - array of 20-byte token addresses (order matches protocol convention) • asset_managers BYTEA[] NULL - per-token managers (e.g., Balancer) • amp NUMERIC(6) NULL - amplification for stable/weighted math • pool_type TEXT NULL - optional human-readable type label • weights NUMERIC(6,5)[] NULL - per-token weights in 0..1 (5 dp) • tick_spacing SMALLINT NULL - grid size for concentrated liquidity • _tracing_id BYTEA - deterministic row-level hash • _parent_tracing_ids BYTEA[] - hash(es) of immediate parent rows in the derivation graph • _genesis_tracing_ids BYTEA[] - hash(es) of original sources (genesis of the derivation path) • _created_at TIMESTAMPTZ - Record creation timestamp. • _updated_at TIMESTAMPTZ - Record last update timestamp

liquidity_pool_fee_terms (versioned, non-overlapping) • pool_uid BIGINT NOT NULL - FK → liquidity_pools(uid) • block_number BIGINT - first block where the token was recognized • block_time TIMESTAMPTZ - UTC timestamp when the block was mined • tx_index INTEGER - tx index for that event • log_index INTEGER - log index for that event • pool_fee NUMERIC(18,18) NOT NULL - Total fee fraction (e.g. 0.003 = 0.30 %) • user_fee_bps SMALLINT NULL - Optional user-side fee share (0–10 000) • protocol_fee_bps SMALLINT NULL - Optional protocol-side fee share (0–10 000) • fee_source TEXT NOT NULL - Provenance of fee rate (e.g. onchain:event) • fee_share_source TEXT NOT NULL - Provenance of fee split (e.g. onchain:param, docs) • _tracing_id BYTEA - deterministic row-level hash • _parent_tracing_ids BYTEA[] - hash(es) of immediate parent rows in the derivation graph • _genesis_tracing_ids BYTEA[] - hash(es) of original sources (genesis of the derivation path) • _created_at TIMESTAMPTZ - Record creation timestamp. • _updated_at TIMESTAMPTZ - Record last update timestamp

Checks • At least one identifier present (contract_address or pool_id) and lengths enforced (20B/32B).

Notes • Fee terms are non-overlapping; each record defines a valid block-range. • Use liquidity_pool_fee_terms for historical fee reconstruction or to obtain the active fee at a given block.

Lineage Every row has a verifiable path back to the originating raw events via the lineage triple and tracing graph: • _tracing_id - this row’s identity • _parent_tracing_ids - immediate sources • _genesis_tracing_ids - original on-chain sources This supports audits and exact reprocessing to source transactions/logs/function calls.

Common Use Cases • Building the complete DEX pool registry for routing and analytics • Filtering pools by fee, type, or token pair • Integrating with reserves, price, and swap datasets for liquidity intelligence • MEV routing, arbitrage path optimization, and chain-wide pool analytics • Constructing pool-level AI or quantitative features

Quality • Each row includes a cryptographic hash linking back to raw on-chain events for auditability. • Tick-level resolution for precision. • Reorg-aware ingestion ensuring data integrity. • Complete backfills to chain genesis for consistency.

Dataset Overview The Level-3 Stablecoin dataset provides complete tick-level liquidity grids for all Uniswap V3-style pools containing stablecoin pairs (e.g., USDC/USDT, DAI/USDC). Each snapshot represents the entire liquidity surface − showing the exact executable size, slippage, and price impact across the full tick range (−887 272 → +887 272).

Key traits • Focused on USD and major fiat-pegged tokens (USDC, USDT, DAI, FRAX, LUSD, crvUSD, etc.) • Schema-stable, versioned, and audit-ready • Real-time (WSS) and historical/EOD delivery • Built for deep-liquidity and peg-stability analytics

Chains and Coverage ETH, BSC, Base, Arbitrum, Unichain, Avalanche, Polygon, Celo, Linea, Optimism (others on request). Full history from chain genesis; reorg-aware real-time ingestion and updates. Coverage includes all major DEX protocols holding stablecoin pairs: • Uniswap V2, V3, V4 • Curve, Balancer, Aerodrome, Solidly, Maverick, Pancake, and others

Schema List the columns exactly as delivered. • snapshot_id BIGINT - unique identifier for each grid snapshot. • pool_uid BIGINT - reference to the liquidity pool (liquidity_pools.uid). • block_number BIGINT - block number of the originating event. • tx_index INTEGER - transaction index within that block. • log_index INTEGER - log index within the transaction. • token_in BYTEA - 20-byte ERC-20 address of the input token. • token_out BYTEA - 20-byte ERC-20 address of the output token. • current_price NUMERIC(78,18) - mid-price (token_out per 1 token_in, decimals-adjusted). • grid_step_bps SMALLINT - spacing between grid points, in basis points. • grid_radius_bps INTEGER - total radius of the grid window, in basis points. • grid_points SMALLINT - number of grid points; must equal radius/step + 1. • reserves_in_adj NUMERIC(78,18) - adjusted reserve amount of the input token (decimals-normalized). • reserves_out_adj NUMERIC(78,18) - adjusted reserve amount of the output token (decimals-normalized). • _tracing_id BYTEA - deterministic row-level hash • _parent_tracing_ids BYTEA[] - hash(es) of immediate parent rows in the derivation graph • _genesis_tracing_ids BYTEA[] - hash(es) of original sources (genesis of the derivation path) • _created_at TIMESTAMPTZ - Record creation timestamp. • _updated_at TIMESTAMPTZ - Record last update timestamp

token_to_token_l3_points • snapshot_id BIGINT - reference to the parent snapshot (token_to_token_l3_snapshots.snapshot_id). • point_index SMALLINT - sequential index (0 … grid_points − 1). • offset_bps_abs INTEGER - absolute offset from the mid-price, in basis points. • size_in NUMERIC(78,18) - executable input amount required to reach this offset. • size_out NUMERIC(78,18) - corresponding output amount at that offset. • price_at_point NUMERIC(78,18) - average price (out / in) including impact.

Lineage Every row has a verifiable path back to the originating raw events via the lineage triple and tracing graph: • _tracing_id - this row’s identity • _parent_tracing_ids - immediate sources • _genesis_tracing_ids - original on-chain sources This supports audits and exact reprocessing to source transactions/logs/function calls.

Common Use Cases • Peg-stability and liquidity-depth analysis for stablecoins • Execution cost and slippage modeling at tick-level precision • Liquidity surface visualizations for Uniswap V3/V4 pools • Cross-pool liquidity comparisons and routing research • Quantitative modeling for arbitrage, MEV, and impact forecasting

Quality • Each row includes a cryptographic hash linking back to raw on-chain events for auditability. • Tick-level resolution for precision. • Reorg-aware ingestion ensuring data integrity. • Complete backfills to chain genesis for consistency.

Dataset Overview Each transaction entry reflects the exact canonical form of the transaction as included in a block, including sender/recipient addresses, gas usage, execution status, transaction type, and (when applicable) created contract addresses.

All records include a deterministic _tracing_id that links each row to its genesis source event, enabling full reconstruction of execution flow and downstream derivations.

Chains and Coverage ETH, BSC, Base, Arbitrum, Unichain, Avalanche, Polygon, Celo, Linea, Optimism (others on request). Full history from chain genesis; reorg-aware real-time ingestion and updates.

Schema List of columns exactly as delivered: • tx_hash BYTEA - 32-byte Keccak-256 transaction hash • from_address BYTEA - 20-byte sender address • to_address BYTEA - 20-byte recipient address (NULL for contract creation) • block_number BIGINT - Block number containing the transaction • tx_index INTEGER - Zero-based index of the transaction within the block • created_contract_address BYTEA - 20-byte contract address created by this transaction (if applicable) • gas_used BIGINT - Gas consumed during transaction execution • effective_gas_price_wei NUMERIC(38,0) - Final gas price paid in wei • status_success BOOLEAN - TRUE if successful, FALSE if reverted, NULL for pre-Byzantium blocks • root BYTEA - Post-transaction state root for pre-Byzantium blocks (NULL otherwise) • tx_type SMALLINT - Transaction type (e.g., 1 = legacy, 2 = EIP-1559) • _tracing_id BYTEA - Deterministic lineage identifier for this transaction • _created_at TIMESTAMPTZ - Record creation timestamp • _updated_at TIMESTAMPTZ - Record last update timestamp

Notes • All binary values can be converted to hex via encode(column, 'hex') for readability. • Unique constraint ensures canonical (block_number, tx_index) ordering. • tx_index and block inclusion can be directly verified with canonical RLP.

Lineage Each tx record includes a deterministic _tracing_id, forming the root lineage reference for all derived BlockDB datasets (swaps, liquidity, and token prices). This ensures verifiable traceability, reproducibility, and proof-of-derivation for every downstream record.

Common Use Cases • Ground-truth layer for swaps, liquidity, and token analytics • Backtesting models using canonical transaction flow • Gas market analysis (EIP-1559, congestion modeling, MEV impact) • Contract deployment tracking and behavior analysis

Quality • Verifiable lineage: deterministic cryptographic hashes per row • Reorg-aware ingestion: continuity and consistency across forks • Complete historical coverage: from chain genesis to present

Context:

This dataset is an extra updating dataset for the G-Research Crypto Forecasting competition.

Introduction

This is a daily updated dataset, automaticlly collecting market data for G-Research crypto forecasting competition. The data is of the 1-minute resolution, collected for all competition assets and both retrieval and uploading are fully automated. see discussion topic.

The Data

For every asset in the competition, the following fields from Binance's official API endpoint for historical candlestick data are collected, saved, and processed.

1. **timestamp** - A timestamp for the minute covered by the row.
2. **Asset_ID** - An ID code for the cryptoasset.
3. **Count** - The number of trades that took place this minute.
4. **Open** - The USD price at the beginning of the minute.
5. **High** - The highest USD price during the minute.
6. **Low** - The lowest USD price during the minute.
7. **Close** - The USD price at the end of the minute.
8. **Volume** - The number of cryptoasset u units traded during the minute.
9. **VWAP** - The volume-weighted average price for the minute.
10. **Target** - 15 minute residualized returns. See the 'Prediction and Evaluation section of this notebook for details of how the target is calculated.
11. **Weight** - Weight, defined by the competition hosts [here](https://www.kaggle.com/cstein06/tutorial-to-the-g-research-crypto-competition)
12. **Asset_Name** - Human readable Asset name.

Indexing

The dataframe is indexed by timestamp and sorted from oldest to newest. The first row starts at the first timestamp available on the exchange, which is July 2017 for the longest-running pairs.

Usage Example

The following is a collection of simple starter notebooks for Kaggle's Crypto Comp showing PurgedTimeSeries in use with the collected dataset. Purged TimesSeries is explained here. There are many configuration variables below to allow you to experiment. Use either GPU or TPU. You can control which years are loaded, which neural networks are used, and whether to use feature engineering. You can experiment with different data preprocessing, model architecture, loss, optimizers, and learning rate schedules. The extra datasets contain the full history of the assets in the same format as the competition, so you can input that into your model too.

Baseline Example Notebooks:

• Neural Network Starter
• LightGBM Starter
• Catboost Starter
• XGBoost Starter
• TabNet Starter
• Reinforcement Learning (PPO) Starter

These notebooks follow the ideas presented in my "Initial Thoughts" here. Some code sections have been reused from Chris' great (great) notebook series on SIIM ISIC melanoma detection competition here

Loose-ends:

This is a work in progress and will be updated constantly throughout the competition. At the moment, there are some known issues that still needed to be addressed:

• VWAP: - At the moment VWAP calculation formula is still unclear. Currently the dataset uses an approximation calculated from the Open, High, Low, Close, Volume candlesticks. [Waiting for competition hosts input]
• Target Labeling: There exist some mismatches to the original target provided by the hosts at some time intervals. On all the others - it is the same. The labeling code can be seen here. [Waiting for competition hosts] input]
• Filtering: No filtration of 0 volume data is taken place.

Example Visualisations

Opening price with an added indicator (MA50): https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fb8664e6f26dc84e9a40d5a3d915c9640%2Fdownload.png?generation=1582053879538546&alt=media" alt="">

Volume and number of trades: https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fcd04ed586b08c1576a7b67d163ad9889%2Fdownload-1.png?generation=1582053899082078&alt=media" alt="">

License

This data is being collected automatically from the crypto exchange Binance.

Context:

This dataset is an extra updating dataset for the G-Research Crypto Forecasting competition.

Introduction

This is a daily updated dataset, automaticlly collecting market data for G-Research crypto forecasting competition. The data is of the 1-minute resolution, collected for all competition assets and both retrieval and uploading are fully automated. see discussion topic.

The Data

For every asset in the competition, the following fields from Binance's official API endpoint for historical candlestick data are collected, saved, and processed.

1. **timestamp** - A timestamp for the minute covered by the row.
2. **Asset_ID** - An ID code for the cryptoasset.
3. **Count** - The number of trades that took place this minute.
4. **Open** - The USD price at the beginning of the minute.
5. **High** - The highest USD price during the minute.
6. **Low** - The lowest USD price during the minute.
7. **Close** - The USD price at the end of the minute.
8. **Volume** - The number of cryptoasset u units traded during the minute.
9. **VWAP** - The volume-weighted average price for the minute.
10. **Target** - 15 minute residualized returns. See the 'Prediction and Evaluation section of this notebook for details of how the target is calculated.
11. **Weight** - Weight, defined by the competition hosts [here](https://www.kaggle.com/cstein06/tutorial-to-the-g-research-crypto-competition)
12. **Asset_Name** - Human readable Asset name.

Indexing

The dataframe is indexed by timestamp and sorted from oldest to newest. The first row starts at the first timestamp available on the exchange, which is July 2017 for the longest-running pairs.

Usage Example

The following is a collection of simple starter notebooks for Kaggle's Crypto Comp showing PurgedTimeSeries in use with the collected dataset. Purged TimesSeries is explained here. There are many configuration variables below to allow you to experiment. Use either GPU or TPU. You can control which years are loaded, which neural networks are used, and whether to use feature engineering. You can experiment with different data preprocessing, model architecture, loss, optimizers, and learning rate schedules. The extra datasets contain the full history of the assets in the same format as the competition, so you can input that into your model too.

Baseline Example Notebooks:

• Neural Network Starter
• LightGBM Starter
• Catboost Starter
• XGBoost Starter
• TabNet Starter
• Reinforcement Learning (PPO) Starter

These notebooks follow the ideas presented in my "Initial Thoughts" here. Some code sections have been reused from Chris' great (great) notebook series on SIIM ISIC melanoma detection competition here

Loose-ends:

This is a work in progress and will be updated constantly throughout the competition. At the moment, there are some known issues that still needed to be addressed:

• VWAP: - At the moment VWAP calculation formula is still unclear. Currently the dataset uses an approximation calculated from the Open, High, Low, Close, Volume candlesticks. [Waiting for competition hosts input]
• Target Labeling: There exist some mismatches to the original target provided by the hosts at some time intervals. On all the others - it is the same. The labeling code can be seen here. [Waiting for competition hosts] input]
• Filtering: No filtration of 0 volume data is taken place.

Example Visualisations

Opening price with an added indicator (MA50): https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fb8664e6f26dc84e9a40d5a3d915c9640%2Fdownload.png?generation=1582053879538546&alt=media" alt="">

Volume and number of trades: https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fcd04ed586b08c1576a7b67d163ad9889%2Fdownload-1.png?generation=1582053899082078&alt=media" alt="">

License

This data is being collected automatically from the crypto exchange Binance.

Dataset Overview Canonical on-chain token reference for fungible and non-fungible assets, providing unified structure and lineage for every recognized contract. Each row represents a unique token or collection, traceable to its genesis event and ABI-decoded metadata.

Chains and Coverage ETH, BSC, Base, Arbitrum, Unichain, Avalanche, Polygon, Celo, Linea, Optimism (others on request). Full history from chain genesis; reorg-aware real-time ingestion and updates. Includes both native coins (ETH, BNB, AVAX, etc.) and token contracts (ERC-20, ERC-721, ERC-1155, ERC-4626, custom standards).

Schema List the columns exactly as delivered. • contract_address BYTEA - PK; 20-byte contract address • block_number BIGINT - first block where the token was recognized • block_time TIMESTAMPTZ - UTC timestamp when the block was mined • tx_index INTEGER - tx index for that event • log_index INTEGER - log index for that event • name TEXT - asset name (from ABI or native coin registry) • symbol TEXT - token symbol or ticker • decimals SMALLINT - number of decimal places for fungible tokens (NULL for NFTs) • metadata_uri TEXT - optional field for NFT metadata base URI (if applicable) • _tracing_id BYTEA - deterministic row-level hash • _parent_tracing_ids BYTEA[] - hash(es) of immediate parent rows in the derivation graph • _genesis_tracing_ids BYTEA[] - hash(es) of original sources (genesis of the derivation path) • _created_at TIMESTAMPTZ - Record creation timestamp. • _updated_at TIMESTAMPTZ - Record last update timestamp

Notes • Use encode(contract_address,'hex') for hex presentation. • Metadata for each token type is retrieved deterministically via ABI decoding or registry sources. • If the ABI read was unsuccessful, the token is not present in this table.

Lineage Every row has a verifiable path back to the originating raw events via the lineage triple and tracing graph: • _tracing_id - this row’s identity • _parent_tracing_ids - immediate sources • _genesis_tracing_ids - original on-chain sources This supports audits and exact reprocessing to source transactions/logs/function calls.

Common Use Cases • Canonical token registry for normalization across DeFi datasets • Symbol, name, decimals lookups for accurate unit scaling in analytics • Cross-chain asset identity resolution • Foundation for NFT, LP token, and vault datasets • Integration layer for pricing engines, wallets, and indexers

Quality • Each row includes a cryptographic hash linking back to raw on-chain events for auditability. • Tick-level resolution for precision. • Reorg-aware ingestion ensuring data integrity. • Complete backfills to chain genesis for consistency.

Context:

This dataset is an extra updating dataset for the G-Research Crypto Forecasting competition.

Introduction

This is a daily updated dataset, automaticlly collecting market data for G-Research crypto forecasting competition. The data is of the 1-minute resolution, collected for all competition assets and both retrieval and uploading are fully automated. see discussion topic.

The Data

For every asset in the competition, the following fields from Binance's official API endpoint for historical candlestick data are collected, saved, and processed.

1. **timestamp** - A timestamp for the minute covered by the row.
2. **Asset_ID** - An ID code for the cryptoasset.
3. **Count** - The number of trades that took place this minute.
4. **Open** - The USD price at the beginning of the minute.
5. **High** - The highest USD price during the minute.
6. **Low** - The lowest USD price during the minute.
7. **Close** - The USD price at the end of the minute.
8. **Volume** - The number of cryptoasset u units traded during the minute.
9. **VWAP** - The volume-weighted average price for the minute.
10. **Target** - 15 minute residualized returns. See the 'Prediction and Evaluation section of this notebook for details of how the target is calculated.
11. **Weight** - Weight, defined by the competition hosts [here](https://www.kaggle.com/cstein06/tutorial-to-the-g-research-crypto-competition)
12. **Asset_Name** - Human readable Asset name.

Indexing

The dataframe is indexed by timestamp and sorted from oldest to newest. The first row starts at the first timestamp available on the exchange, which is July 2017 for the longest-running pairs.

Usage Example

The following is a collection of simple starter notebooks for Kaggle's Crypto Comp showing PurgedTimeSeries in use with the collected dataset. Purged TimesSeries is explained here. There are many configuration variables below to allow you to experiment. Use either GPU or TPU. You can control which years are loaded, which neural networks are used, and whether to use feature engineering. You can experiment with different data preprocessing, model architecture, loss, optimizers, and learning rate schedules. The extra datasets contain the full history of the assets in the same format as the competition, so you can input that into your model too.

Baseline Example Notebooks:

• Neural Network Starter
• LightGBM Starter
• Catboost Starter
• XGBoost Starter
• TabNet Starter
• Reinforcement Learning (PPO) Starter

These notebooks follow the ideas presented in my "Initial Thoughts" here. Some code sections have been reused from Chris' great (great) notebook series on SIIM ISIC melanoma detection competition here

Loose-ends:

This is a work in progress and will be updated constantly throughout the competition. At the moment, there are some known issues that still needed to be addressed:

• VWAP: - At the moment VWAP calculation formula is still unclear. Currently the dataset uses an approximation calculated from the Open, High, Low, Close, Volume candlesticks. [Waiting for competition hosts input]
• Target Labeling: There exist some mismatches to the original target provided by the hosts at some time intervals. On all the others - it is the same. The labeling code can be seen here. [Waiting for competition hosts] input]
• Filtering: No filtration of 0 volume data is taken place.

Example Visualisations

Opening price with an added indicator (MA50): https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fb8664e6f26dc84e9a40d5a3d915c9640%2Fdownload.png?generation=1582053879538546&alt=media" alt="">

Volume and number of trades: https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fcd04ed586b08c1576a7b67d163ad9889%2Fdownload-1.png?generation=1582053899082078&alt=media" alt="">

License

This data is being collected automatically from the crypto exchange Binance.

Dataset Overview Canonical ERC-20 token reference with deterministic tracing at the row level. One row per token contract, with audit-grade lineage to the first recognition event and to parent/genesis derivations.

Chains and Coverage ETH, BSC, Base, Arbitrum, Unichain, Avalanche, Polygon, Celo, Linea, Optimism (others on request). Full history from chain genesis; reorg-aware real-time ingestion and updates.

Schema List the columns exactly as delivered. • contract_address BYTEA - PK; 20-byte ERC-20 contract address • block_number BIGINT - first block where the token was recognized • block_time TIMESTAMPTZ - UTC timestamp when the block was mined • tx_index INTEGER - tx index for that event • log_index INTEGER - log index for that event • name TEXT - asset name • symbol TEXT - token symbol or ticker • decimals SMALLINT - number of decimal places • _tracing_id BYTEA - deterministic row-level hash • _parent_tracing_ids BYTEA[] - hash(es) of immediate parent rows in the derivation graph • _genesis_tracing_ids BYTEA[] - hash(es) of original sources (genesis of the derivation path) • _created_at TIMESTAMPTZ - Record creation timestamp. • _updated_at TIMESTAMPTZ - Record last update timestamp

Notes • Use encode(contract_address,'hex') for hex presentation. • Metadata for each token type is retrieved deterministically via ABI decoding or registry sources. • If the ABI read was unsuccessful, the token is not present in this table.

Lineage Every row has a verifiable path back to the originating raw events via the lineage triple and tracing graph: • _tracing_id - this row’s identity • _parent_tracing_ids - immediate sources • _genesis_tracing_ids - original on-chain sources This supports audits and exact reprocessing to source transactions/logs/function calls.

Common Use Cases • Token registry to normalize joins for swaps, transfers, pools, and prices • Amount scaling via decimals for analytics, PnL, and model features • App backends: display names/symbols and validate token addresses

Quality • Each row includes a cryptographic hash linking back to raw on-chain events for auditability. • Tick-level resolution for precision. • Reorg-aware ingestion ensuring data integrity. • Complete backfills to chain genesis for consistency.

Context:

This dataset is an extra updating dataset for the G-Research Crypto Forecasting competition.

Introduction

This is a daily updated dataset, automaticlly collecting market data for G-Research crypto forecasting competition. The data is of the 1-minute resolution, collected for all competition assets and both retrieval and uploading are fully automated. see discussion topic.

The Data

For every asset in the competition, the following fields from Binance's official API endpoint for historical candlestick data are collected, saved, and processed.

1. **timestamp** - A timestamp for the minute covered by the row.
2. **Asset_ID** - An ID code for the cryptoasset.
3. **Count** - The number of trades that took place this minute.
4. **Open** - The USD price at the beginning of the minute.
5. **High** - The highest USD price during the minute.
6. **Low** - The lowest USD price during the minute.
7. **Close** - The USD price at the end of the minute.
8. **Volume** - The number of cryptoasset u units traded during the minute.
9. **VWAP** - The volume-weighted average price for the minute.
10. **Target** - 15 minute residualized returns. See the 'Prediction and Evaluation section of this notebook for details of how the target is calculated.
11. **Weight** - Weight, defined by the competition hosts [here](https://www.kaggle.com/cstein06/tutorial-to-the-g-research-crypto-competition)
12. **Asset_Name** - Human readable Asset name.

Indexing

The dataframe is indexed by timestamp and sorted from oldest to newest. The first row starts at the first timestamp available on the exchange, which is July 2017 for the longest-running pairs.

Usage Example

The following is a collection of simple starter notebooks for Kaggle's Crypto Comp showing PurgedTimeSeries in use with the collected dataset. Purged TimesSeries is explained here. There are many configuration variables below to allow you to experiment. Use either GPU or TPU. You can control which years are loaded, which neural networks are used, and whether to use feature engineering. You can experiment with different data preprocessing, model architecture, loss, optimizers, and learning rate schedules. The extra datasets contain the full history of the assets in the same format as the competition, so you can input that into your model too.

Baseline Example Notebooks:

• Neural Network Starter
• LightGBM Starter
• Catboost Starter
• XGBoost Starter
• TabNet Starter
• Reinforcement Learning (PPO) Starter

These notebooks follow the ideas presented in my "Initial Thoughts" here. Some code sections have been reused from Chris' great (great) notebook series on SIIM ISIC melanoma detection competition here

Loose-ends:

This is a work in progress and will be updated constantly throughout the competition. At the moment, there are some known issues that still needed to be addressed:

• VWAP: - At the moment VWAP calculation formula is still unclear. Currently the dataset uses an approximation calculated from the Open, High, Low, Close, Volume candlesticks. [Waiting for competition hosts input]
• Target Labeling: There exist some mismatches to the original target provided by the hosts at some time intervals. On all the others - it is the same. The labeling code can be seen here. [Waiting for competition hosts] input]
• Filtering: No filtration of 0 volume data is taken place.

Example Visualisations

Opening price with an added indicator (MA50): https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fb8664e6f26dc84e9a40d5a3d915c9640%2Fdownload.png?generation=1582053879538546&alt=media" alt="">

Volume and number of trades: https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2234678%2Fcd04ed586b08c1576a7b67d163ad9889%2Fdownload-1.png?generation=1582053899082078&alt=media" alt="">

License

This data is being collected automatically from the crypto exchange Binance.

Dataset Overview The Level-2 (L2) stablecoin liquidity impact dataset provides quantized liquidity grids around the mid-price (typically ±10%) for all major USD-pegged stablecoin pairs. Each row represents a depth snapshot at a defined basis-point spacing (e.g. 10 bps), describing how much stablecoin can be executed before the price deviates by a certain percentage.

These grids allow analysts and developers to model depth curves, slippage, and peg resilience with high precision, fully reproducible from on-chain data.

Key traits • Focused on USD and major fiat-pegged tokens (USDC, USDT, DAI, FRAX, LUSD, crvUSD, etc.) • Schema-stable, versioned, and audit-ready • Real-time (WSS) and historical/EOD delivery • Fully verifiable lineage back to pool events and raw on-chain logs

Chains and Coverage ETH, BSC, Base, Arbitrum, Unichain, Avalanche, Polygon, Celo, Linea, Optimism (others on request). Full history from chain genesis; reorg-aware real-time ingestion and updates. Coverage includes all major DEX protocols holding stablecoin pairs: • Uniswap V2, V3, V4 • Curve, Balancer, Aerodrome, Solidly, Maverick, Pancake, and others

Schema List the columns exactly as delivered. • snapshot_id BIGINT - unique identifier for each grid snapshot. • pool_uid BIGINT - reference to the liquidity pool (liquidity_pools.uid). • block_number BIGINT - block number of the originating event. • tx_index INTEGER - transaction index within that block. • log_index INTEGER - log index within the transaction. • token_in BYTEA - 20-byte ERC-20 address of the input token. • token_out BYTEA - 20-byte ERC-20 address of the output token. • current_price NUMERIC(78,18) - mid-price (token_out per 1 token_in, decimals-adjusted). • grid_step_bps SMALLINT - spacing between grid points, in basis points. • grid_radius_bps INTEGER - total radius of the grid window, in basis points. • grid_points SMALLINT - number of grid points; must equal radius/step + 1. • _tracing_id BYTEA - deterministic row-level hash • _parent_tracing_ids BYTEA[] - hash(es) of immediate parent rows in the derivation graph • _genesis_tracing_ids BYTEA[] - hash(es) of original sources (genesis of the derivation path) • _created_at TIMESTAMPTZ - Record creation timestamp. • _updated_at TIMESTAMPTZ - Record last update timestamp

token_to_token_l2_points • snapshot_id BIGINT - reference to the parent snapshot (token_to_token_l2_snapshots.snapshot_id). • point_index SMALLINT - sequential index (0 … grid_points − 1). • offset_bps_abs INTEGER - absolute offset from the mid-price, in basis points. • size_in NUMERIC(78,18) - executable input amount required to reach this offset. • size_out NUMERIC(78,18) - corresponding output amount at that offset. • price_at_point NUMERIC(78,18) - average price (out / in) including impact.

(For hex display: encode(token_in,'hex'), encode(token_out,'hex').)

Lineage Every row has a verifiable path back to the originating raw events via the lineage triple and tracing graph: • _tracing_id - this row’s identity • _parent_tracing_ids - immediate sources • _genesis_tracing_ids - original on-chain sources This supports audits and exact reprocessing to source transactions/logs/function calls.

Common Use Cases • Stablecoin peg analysis through liquidity depth modeling • Execution algorithm calibration (impact-aware order sizing) • Market-making and slippage profiling • Cross-chain peg risk and liquidity fragmentation studies • MEV and arbitrage signal extraction • AI/ML feature generation for depeg and volatility prediction

Quality • Each row includes a cryptographic hash linking back to raw on-chain events for auditability. • Tick-level resolution for precision. • Reorg-aware ingestion ensuring data integrity. • Complete backfills to chain genesis for consistency.