RESEARCH_DATASET // RELEASE_2.0.0

BUY THE FORM 4
RESEARCH RELEASE.

A one-time SEC Form 4 research package with 1.63M publishable rows, CSV/Parquet/DuckDB delivery, a free 10,000-row sample and commercial-license contact for team or product use.

Founding buyer price

InsiderAlpha Form 4 Research Database

$99one-time payment

Regular price $199

1,633,946 transactions
CSV + Parquet + DuckDB
Internal research license
Instant download
3 months of API access included

Buy full dataset — $99

Download complete sample

Need team or commercial use? Contact us

1,633,946

Publishable rows

Documented fields

2.0.0

Schema version

1.5 GB

Uncompressed CSV

Evaluate before purchase

Inspect 10,000 real rows and the complete 64-field schema.

The free sample uses deterministic positions from a versioned schema 2.0.0 release. No email gate, no synthetic rows and no hidden columns. The ZIP includes a manifest, methodology note and SHA-256 checksums.

2.1 MBZIP · UTF-8 CSVNo executables or macros

Download complete sample

Product proof

Start researching in minutes, not weeks

Query the complete history immediately with DuckDB.
Reproduce the included insider-purchase studies.
Filter purchases, sales, roles and clusters with reusable SQL.
Avoid collecting, parsing and normalizing 1.63M SEC records.

Ready-to-query database

Open the packaged database and query a convenience view without importing the 1.5 GB CSV.

SELECT *
FROM open_market_purchases
ORDER BY "Date" DESC
LIMIT 20;

Guided notebook included

Start from a Jupyter workflow, reusable SQL and documented views for purchases, sales and cluster analysis.

Inspect the 64-field contract

Published research outputs

See the type of event studies the release supports before buying the complete research package.

Research contract

One table. Traceable inputs. Reproducible interpretation.

The export is deliberately more than a list of names and transaction values. Canonical identity fields make SEC rows auditable; quality columns describe how each row entered the publishable set; derived columns expose the exact enrichment used in the product.

SEC filing lineage, accession number and canonical row identity
Transaction code, shares, price, ownership and nominal value
Quality, validation and source-verification states
Issuer sector, industry, SIC and metadata provenance
Normalized executive roles and reporting-owner CIK
7D, 30D and 90D purchase-cluster indicators
Security and insider-basis outcomes at 30D, 90D and 180D
Equivalent CSV, Parquet and ready-to-query DuckDB files
Jupyter quickstart notebook, reusable SQL and offline documentation
Release manifest with row count, schema version and SHA-256

Interactive schema preview

Inspect the fields before you commit

Full 64-field dictionary

Switch between source, lineage, quality and enrichment groups. Every field exposes its type, origin, nullability and a representative value; examples describe the schema and are not a live customer record.

Core transaction fields

Stable compatibility fields available in API transaction responses and the bulk release.

Field	Type	Origin	Example
`id` Stable internal transaction identifier.	integerRequired	SEC / legacy source	`1568928`
`Form` Legacy-compatible display label for the filing form.	stringRequired	SEC / legacy source	`Form 4`
`Ticker` Normalized uppercase issuer ticker.	stringRequired	SEC / legacy source	`AAPL`
`Executive` Reporting-owner name as stored by the transaction pipeline.	stringRequired	SEC / legacy source	`DOE JOHN`
`Position` Reported relationship or officer title before standardized role enrichment.	stringNullable	SEC / legacy source	`Chief Financial Officer`
`Date` Transaction date. This is not necessarily the filing or acceptance date.	dateRequired	SEC / legacy source	`2026-07-23`
`Type` SEC transaction code such as P, S, A, M, F or G.	stringRequired	SEC / legacy source	`P`
`Shares` Number of securities reported in the transaction.	numberRequired	SEC / legacy source	`500`
`Price` Reported per-security transaction price. Zero can be valid for grants and other technical events.	numberRequired	SEC / legacy source	`202.45`
`Total_Value` Legacy-compatible nominal value stored with the transaction.	numberRequired	SEC / legacy source	`101225.00`
`Notes` Footnote or explanatory filing text when available.	stringNullable	SEC / legacy source	`Shares sold pursuant to a Rule 10b5-1 plan.`

Reproducible starting points

Three SQL queries you can run immediately

These queries run against the packaged DuckDB database and use the included convenience views. Copy them as a starting point, then inspect the methodology before interpreting any result as a signal.

Find multi-insider purchase clusters

Rank issuers with at least three distinct Code P buyers in the trailing 90 days.

SELECT
    "Ticker",
    COUNT(*) AS purchases,
    COUNT(DISTINCT "Executive") AS distinct_buyers,
    SUM(calculated_total_value) AS disclosed_value
FROM open_market_purchases
WHERE "Date" >= CURRENT_DATE - INTERVAL 90 DAY
GROUP BY "Ticker"
HAVING COUNT(DISTINCT "Executive") >= 3
ORDER BY distinct_buyers DESC, disclosed_value DESC
LIMIT 50;

Compare purchases by executive role

Build a role cohort using calculated 30-day security outcomes and a minimum sample size.

SELECT
    standard_role,
    COUNT(*) AS purchases,
    MEDIAN(market_return_30d) AS median_market_return_30d
FROM open_market_purchases
WHERE return_status_30d = 'calculated'
GROUP BY standard_role
HAVING COUNT(*) >= 30
ORDER BY purchases DESC;

Audit provenance before analysis

Measure the exact mix of canonical SEC rows and validated historical records in the release.

SELECT
    validation_status,
    source_verification_status,
    COUNT(*) AS rows,
    ROUND(100.0 * COUNT(*) / SUM(COUNT(*)) OVER (), 2) AS share_pct
FROM transactions
GROUP BY validation_status, source_verification_status
ORDER BY rows DESC;

Validate the workflow

Test the schema before running the complete queries

Import the free sample, inspect column types and confirm how identifiers, null values and transaction codes behave in your own stack.

Download complete sample

Current release

Coverage, disclosed—not implied

Inspect all 64 fields

806.6K

SEC verified · 49.4%

Canonical Form 4/4-A transactions identified by accession, transaction table and sequence.

827.4K

Validated legacy · 50.6%

Historical rows retained after deterministic quality rules; their status remains visible instead of being presented as canonical SEC parses.

Sector & industry1.2M populated74.0% coverage · 26.0% null

Standardized role1.6M populated100.0% coverage · 0.0% null

30D market outcome1.5M populated94.1% coverage · 5.9% null

90D market outcome1.5M populated92.8% coverage · 7.2% null

180D market outcome1.5M populated88.8% coverage · 11.2% null

Eligibility

Only current rows in sec_verified or validated_legacy state enter the commercial release.

Lineage

Canonical rows preserve accession, filing date, table and transaction sequence when available.

Derived fields

Roles, clusters and price outcomes expose their algorithm or source version in the schema.

Applications

Built for questions, not just downloads

Hedge funds & asset managers

Screen purchase clusters, compare executive cohorts and build auditable event-study inputs without maintaining a Form 4 parser.

Independent researchers

Test transaction code, role, ownership and post-transaction outcomes from a local DuckDB or Parquet workflow.

Fintech teams

Prototype screens and internal tools with the bulk release, then move recurring targeted queries to the documented API.

Universities & labs

Teach or reproduce financial-data research from a versioned export with lineage, quality states and explicit limitations.

Choose the delivery model

Sample, complete release or API

The products solve different jobs. Use the sample to validate fit, the complete release for exhaustive local research, and the API for targeted application requests.

Evaluate

Free sample

Validate the schema and import workflow with a deterministic subset.

Best for
Schema review and technical evaluation

10,000 real rows
All 64 fields
CSV + manifest
No email gate

Download complete sample

Research

Complete dataset · $99

Run full-history research locally from one versioned snapshot.

Founding buyer price · one-time payment · regular $199

Best for
Backtests, joins, event studies and academic work

1,633,946 rows
CSV · Parquet · DuckDB
Internal research license
3 months of API access

Buy full dataset — $99

Integrate

RapidAPI access

Request targeted JSON from applications and recurring workflows.

Best for
Fintech products, dashboards and operational queries

Eight REST endpoints
Ticker and executive lookup
Windowed analytics
RapidAPI billing and keys

View API plans Compare API access

Dataset checkout terms are controlled by Gumroad. API pricing, quotas, overages and renewal terms are controlled by the plan displayed on RapidAPI at the time you subscribe.

Buyer package

Start querying, not converting files

Every format is generated from the same validated release and checked against the same row count. Choose the workflow that fits your research environment.

CSV

Portable UTF-8 source export for databases, scripts and archival workflows.

Parquet

Compressed columnar data for fast analytical scans in DuckDB, Polars, pandas and Spark.

DuckDB

A local database with the full table and convenience views for purchases, sales and clusters.

Notebook + docs

A guided Jupyter analysis, reusable SQL, 64-field dictionary, methodology and integrity checks.

Release history

A versioned product, not a moving download

Current · v2.0.0

Schema 2.0.0 · Commercial release

2026-07-29 UTC · 1,633,946 rows

Current

64-field contract spanning source, SEC lineage, quality, issuer, roles, clusters and price outcomes.
Equivalent CSV, Parquet and DuckDB materializations checked against the same row count.
Convenience views, Jupyter quickstart and reusable SQL included in the buyer archive.
Release manifest and SHA-256 checksums added for artifact-level verification.

This is the first commercially documented release in the public changelog. Future releases are listed here only after their manifests are generated; purchase does not imply future updates unless the checkout explicitly includes them.

Founding customer offer

Data now, API access and a concrete voice in the roadmap

3 months

API access, manually activated after purchase.

5,000/mo

Requests per calendar month for evaluation and internal research.

3 vote points

Allocate them 3, 2 + 1, or 1 + 1 + 1 across candidate features.

Roadmap votes are advisory and influence prioritization; they do not guarantee implementation or timing. API access is non-transferable, subject to fair use and available while this founding offer is displayed.

Methodology guardrails

What the derived fields do—and do not—mean

Directional transactions

Open-market flow treats code P as purchase and code S as sale. Awards, options, gifts and tax withholdings remain available but are not interchangeable directional signals.

Purchase clusters

Cluster counts use distinct reporting owners with code P in inclusive 7, 30 and 90 calendar-day windows. A cluster flag requires at least three owners in the 30-day window.

Post-transaction outcomes

Market outcomes measure the security's price change from the transaction-date close. They are not benchmark adjusted and do not establish causation or predict future performance.

Buyer questions

Before you acquire the release

What exactly is included?

The same versioned research snapshot in CSV, compressed Parquet and ready-to-query DuckDB formats, containing 1,633,946 publishable transactions and 64 documented columns. It also includes a Jupyter quickstart, reusable SQL, an offline data dictionary, methodology, license summary, release manifest and SHA-256 checksums.

Can I evaluate the schema before purchasing?

Yes. The free evaluation archive contains 10,000 real rows and all 64 columns from a schema 2.0.0 release snapshot. Rows are selected at deterministic, evenly spaced positions across that release. The archive also includes its methodology, manifest and checksums.

Is the sample safe to download?

The archive contains only CSV, JSON and TXT files—no executable files, scripts, macros or spreadsheet workbooks. Its build validates the ZIP structure and rejects cells that look like spreadsheet formulas. A published SHA-256 checksum lets you verify that the download was not altered. No download can carry an absolute zero-risk guarantee, so use your normal endpoint-security controls.

Are the return fields investment performance claims?

No. They are descriptive post-transaction price outcomes. Market returns use the security close around the transaction date and do not subtract SPY or another benchmark. Insider-basis returns use the reported transaction price when available.

Is every field complete for every row?

No. SEC records and historical sources vary. Missing values remain null rather than being invented. Current release coverage is published below and the dictionary states when each field can be absent.

Should I open the file in Excel?

The 1.5 GB CSV exceeds the practical row and memory limits of ordinary spreadsheets. The included DuckDB and Parquet files are designed to make the complete release usable without loading it into Excel or into memory.

What does the founding-customer offer include?

While the offer is displayed, a verified buyer can request three months of API access with up to 5,000 requests per calendar month and receives three Founding Customer Roadmap Vote points. Activation is manual through client@insider-alpha.com. Votes influence prioritization but do not guarantee implementation or delivery dates.

Does purchase include future updates or redistribution rights?

The checkout listing controls the exact update entitlement. The standard research license does not permit resale, public redistribution or embedding the raw dataset in another commercial data product. Contact us for commercial licensing.

Release integrity

Every release records its schema version, generation timestamp, row count and checksum. The current SHA-256 begins 002f598df06855ee….

Generated 2026-07-29T21:18:38.313847Z. Research data may contain issuer-reported errors, amendments, null fields and corporate-action edge cases; review the documented limitations before production use.

Read limitations

Data-only archive

The sample contains exactly one CSV, one JSON manifest and two TXT documents. No executables, scripts, macros or workbooks.

Build-time validation

Generation verifies 64 columns, row counts, archive CRC and rejects values that look like spreadsheet formulas.

Verifiable integrity

Confirm the downloaded artifact against its public SHA-256:

850f007868b1c3200f76c986dee8f3013efdff1269a3b3573716aed233a1f0d7

Inspect sample manifest Open checksum file