What Does a Crypto Data Analyst Do?

What does a crypto data analyst do?

A crypto data analyst defines metrics, writes SQL against decoded blockchain tables, builds dashboards, and explains what the numbers mean. The job differs from a classic data analyst role in one way: most source data is public, raw, and encoded, so you must understand how contracts emit events before any metric is correct.

The work falls into five groups:

The shares above are typical patterns, not survey data. They vary by employer: an exchange analyst does more internal reporting, while a protocol analyst does more incentive and governance work.

What does a typical day look like?

A typical day starts with a check of the KPI dashboard for anomalies, such as a volume spike from one wallet. Next come requests from growth or BD, for example the retention of wallets from a partner integration. The afternoon often goes to a longer analysis, such as a cohort study or a sybil review before an airdrop.

What skills does an onchain data analyst need?

An onchain data analyst needs strong SQL, a mental model of how blockchains store data, and the judgment to define metrics that resist manipulation. The most important technical skills are joins and window functions on decoded tables, event decoding, token decimal and price handling, and wallet labeling. Use Python for data that SQL platforms do not cover.

How is Solana data different from EVM data?

EVM chains emit events as logs, with indexed topics and an ABI that decodes them. Solana programs do not emit logs in the same form. You read instructions, inner instructions, and account state changes. On Solana, decoded tables from a data provider save you a lot of time. Raw instruction data is hard to parse by hand.

What tools do crypto data analysts use?

Most crypto data analysts use a public or commercial query platform for chain data, an indexer for custom contracts, a modeling tool, Python for gaps, and a product analytics tool for app behavior. The exact mix depends on budget and on whether the employer runs its own warehouse.

Formo details, as documented: the SQL Explorer and the POST /v0/query API run read-only ClickHouse SQL on your own project data; the @formo/cli CLI runs queries with formo query run; the MCP server lets AI assistants query the same data and requires a Scale or Enterprise plan.

A note on Flipside: in May 2026 Flipside sold its blockchain data business to SonarX and moved its focus to an enterprise AI platform. Check the current status of any Flipside dashboards before you depend on them.

What questions do crypto data analysts answer?

Crypto data analysts answer questions about who uses the protocol, where they came from, whether they stay, and what incentives cost. Growth teams ask about channels and funnels. BD teams ask about integrations and partners. Governance asks about emissions, treasury, and voter activity. Leadership asks for a short, trusted set of KPIs.

What do example analyses look like?

Example analyses start with a clear question, a short SQL query on decoded tables, and a caveat about what the query cannot see. The snippets below use Dune's public tables and DuneSQL syntax, except the last one, which uses ClickHouse SQL on Formo data. Table names were checked against the Dune docs on 2026-11-08.

How do you count daily active traders on a DEX?

-- DuneSQL (Dune). Daily traders and volume for one DEX on Ethereum.
SELECT
  date_trunc('day', block_time) AS day,
  count(DISTINCT tx_from)       AS traders,
  sum(amount_usd)               AS volume_usd
FROM dex.trades
WHERE blockchain = 'ethereum'
  AND project = 'uniswap'
  AND block_time >= now() - interval '30' day
GROUP BY 1
ORDER BY 1

tx_from is the address that sent the transaction. Bots and aggregators inflate it, so label them before you report traders as users.

How do you handle token decimals and prices in SQL?

-- DuneSQL (Dune). Decode raw ERC-20 Transfer logs for USDC and price them per hour.
SELECT
  date_trunc('hour', l.block_time) AS hour,
  sum(
    CAST(varbinary_to_uint256(l.data) AS double)
    / power(10, p.decimals) * p.price
  ) AS usd_transferred
FROM ethereum.logs AS l
JOIN prices.hour AS p
  ON  p.blockchain = 'ethereum'
  AND p.contract_address = l.contract_address
  AND p.timestamp = date_trunc('hour', l.block_time)
WHERE l.contract_address = 0xa0b86991c6218b36c1d19d4a2e9eb0ce3606eb48
  AND l.topic0 = 0xddf252ad1be2c89b69c2b068fc378daa952ba7f163c4a11628f55a4df523b3ef
  AND l.block_time >= now() - interval '1' day
GROUP BY 1
ORDER BY 1

The topic0 value is the keccak-256 hash of the ERC-20 Transfer(address,address,uint256) event signature. The raw amount is in data, so it must be divided by 10^decimals. In practice, curated tables such as tokens_evm.transfers already do this, but you must know how to check them. One example: Dune prices amount_usd in that table from prices_external.hour, not from the prices.hour table used above, so the two totals can differ.

How do you build a weekly retention cohort?

-- DuneSQL (Dune). Share of first-week traders who traded again in later weeks.
WITH first_week AS (
  SELECT tx_from AS wallet, min(date_trunc('week', block_time)) AS cohort
  FROM dex.trades
  WHERE blockchain = 'ethereum' AND project = 'uniswap'
    AND block_time >= now() - interval '90' day
  GROUP BY 1
),
activity AS (
  SELECT DISTINCT tx_from AS wallet, date_trunc('week', block_time) AS week
  FROM dex.trades
  WHERE blockchain = 'ethereum' AND project = 'uniswap'
    AND block_time >= now() - interval '90' day
)
SELECT
  f.cohort,
  date_diff('week', f.cohort, a.week) AS weeks_since_first,
  count(DISTINCT a.wallet)            AS active_wallets
FROM first_week AS f
JOIN activity AS a ON a.wallet = f.wallet
GROUP BY 1, 2
ORDER BY 1, 2

How do you measure the app side of the funnel?

Onchain tables do not show visits, referrers, or wallet connects that never led to a transaction. Product analytics fills that gap. This query uses ClickHouse SQL on the Formo events table, as in the Formo SQL Explorer guide.

-- ClickHouse SQL (Formo Explorer or query API). Daily visitors and wallet connects, by UTC day.
SELECT
  toDate(timestamp)                                     AS date,
  count(DISTINCT if(type = 'page', anonymous_id, NULL)) AS visitors,
  count(DISTINCT if(type = 'connect', address, NULL))   AS connected_wallets
FROM events
WHERE timestamp >= now() - INTERVAL 30 DAY
GROUP BY date
ORDER BY

The visitors column counts anonymous IDs that sent a page view, which is how Formo defines Visitors. Formo days are UTC calendar days. The query shows only your own app: it cannot see wallets that use your contracts through other frontends, so pair it with the onchain queries above.

How much does a crypto data analyst earn?

Public salary data is thin, so treat any figure as a rough guide. Unchain Data's August 2026 update (checked November 8, 2026) reports a median posted pay band of $126,371 for crypto data analysts, from 26 verified postings, with a middle half (P25 to P75) of about $104k to $151k. These are posted bands, not actual pay.

The source uses only postings that publish both a minimum and a maximum, in USD, and takes the midpoint of each band. It includes active and expired postings with no rolling time window, and it refreshes quarterly. Remote roles, token compensation, and location change real pay a lot.

How do you become a crypto data analyst?

You become a crypto data analyst by learning SQL, learning how blockchain data is stored, and publishing analyses that answer real protocol questions. Hiring managers in crypto often look at public dashboards and write-ups before a CV. Plan for a few months of steady practice if you already know some SQL.

1. Learn SQL fundamentals. Practice SELECT, filters, aggregation, joins, CTEs, and window functions.

2. Learn the data model. Study blocks, transactions, logs, traces, event signatures, and token decimals on EVM chains, then Solana instructions and accounts.

3. Practice on real tables. Use the Dune docs and public tables such as dex.trades and prices.hour. Courses help: Unchain Data's WizardCamp has SQL exercises on Ethereum and Solana data, with free exercises and a paid tier.

4. Pick one protocol and go deep. Read its contracts and docs, then rebuild one of its published metrics and reconcile your number with theirs.

5. Publish three analyses. Each one should answer a question a growth, BD, or governance lead would ask, with a chart, the query, and caveats.

6. Add modeling and Python. Learn dbt for tested models and Python for API data. This also opens the path to analytics engineering.

7. Apply where your work is visible. Share your work in protocol forums and governance threads, and apply to protocols, exchanges, and data providers.

If you want to move toward pipelines later, Unchain Data's guide on how to become a crypto data engineer lists a six-step roadmap.

What is the career path for a crypto data analyst?

The usual path goes from analyst to senior analyst, then splits. One branch goes to analytics engineering or data engineering. The other goes to data leadership or a strategy role close to growth or tokenomics.

Frequently asked questions

Do I need to know how to code to be a crypto data analyst?

You need SQL. Most of the daily work happens in SQL on decoded tables. Python is a strong second skill for API data and statistics. You do not need to write smart contracts, but you must be able to read an ABI and understand which events a contract emits.

What is the difference between a crypto data analyst and a crypto data engineer?

A crypto data analyst answers questions and builds dashboards with existing tables. A crypto data engineer builds and maintains the pipelines and models that create those tables, such as ingestion from nodes or indexers, decoding, and warehouse models. Many analysts move into analytics engineering, which sits between the two.

Which SQL dialect does Dune use?

Dune uses DuneSQL, which is based on Trino. Functions for binary data, such as varbinary_to_uint256, and address literals written as 0x values are specific to it. Other platforms use their own dialects, for example ClickHouse SQL in Formo, so check functions when you move a query.

Is Flipside still a good place to learn?

In May 2026 Flipside sold its blockchain data business to SonarX and moved its focus to an enterprise AI platform. Old Flipside dashboards and tutorials still teach useful ideas, but check whether a tool is still active before you build a portfolio on it. Dune and WizardCamp are current alternatives.

How much does a crypto data analyst earn?

Unchain Data's August 2026 update reports a median posted pay band of $126,371 for crypto data analysts, based on 26 verified postings. The middle half falls between about $104k and $151k. These figures are posted bands, so actual offers differ by location, seniority, and token compensation.

Can I become a crypto data analyst with no experience?

Yes, if you can show your work. Learn SQL, practice on public decoded tables, and publish analyses that answer real protocol questions. A public portfolio of dashboards and short write-ups often matters more to crypto hiring managers than past job titles.

Related articles

• The crypto data analytics landscape in 2026

• DeFi KPIs and metrics for growth

• DeFi cohort analysis

• Wallet analytics: the ultimate guide for crypto teams

• Unify onchain and offchain data

• Find at-risk wallets in DeFi