---

Why This Project?

A single investor buying a stock means nothing — they have their own context, biases, and constraints. But when dozens of legendary investors independently buy the same stock, that's a powerful signal.

DataRoma tracks 82 superinvestors' portfolios from SEC 13F filings. This library gives you programmatic access to their consensus — what they're buying together, selling together, and holding together.

from superinvestor import SI

si = SI()
si.buys(n=5)
# [{'symbol': 'AMZN', 'name': 'Amazon.com Inc.', 'buy_count': 11, 'avg_hold_price': 230.82},
#  {'symbol': 'META', 'name': 'Meta Platforms Inc.', 'buy_count': 10, 'avg_hold_price': 660.09},
#  {'symbol': 'RKT', 'name': 'Rocket Companies Inc.', 'buy_count': 6, 'avg_hold_price': 19.36}, ...]

No API key needed. No config. Just pip install and go.

---

Installation

pip install superinvestor

---

Quick Start

from superinvestor import SI

si = SI()

# What are superinvestors buying this quarter?
si.buys(n=5)

# What are they selling?
si.sells(n=5)

# Deep dive into a specific stock
si.stock("AAPL")

# Full consensus portfolio
si.holdings(n=10)

# List all 82 tracked superinvestors
si.managers()

---

API Reference

SI()

Create a client instance. No arguments needed.

.holdings(n=None) → list[dict]

Grand portfolio — all stocks held by superinvestors, ranked by ownership count.

si.holdings(n=3)
# [{'symbol': 'AMZN', 'name': 'Amazon.com Inc.', 'ownership_count': 31,
#   'max_pct': 33.10, 'avg_hold_price': 230.82}, ...]

Field	Description
symbol	Ticker symbol
name	Company name
ownership_count	Number of superinvestors holding this stock
max_pct	Maximum portfolio allocation among holders (%)
avg_hold_price	Weighted average hold price from 13F filings ($)

.buys(period="q", n=None) → list[dict]

Stocks being bought by multiple superinvestors.

Parameter	Type	Description
period	str	"q" for quarterly (default), "6m" for 6 months
n	int	Top N results. None for all

si.buys(period="6m", n=10)

.sells(period="q", n=None) → list[dict]

Stocks being sold by multiple superinvestors. Same parameters as .buys().

.stock(symbol) → dict

Deep dive into a specific stock: who holds it, quarterly activity trends, and consensus direction.

si.stock("AAPL")
# {
#   'symbol': 'AAPL',
#   'sector': 'Technology',
#   'ownership_count': 18,
#   'ownership_rank': 9,
#   'avg_hold_price': 271.86,
#   'quarterly_activity': [
#     {'period': 'Q4 2025',
#      'buy':    {'count': 0, 'shares': 0},
#      'add':    {'count': 2, 'shares': 30100},
#      'reduce': {'count': 11, 'shares': 13140548},
#      'sell':   {'count': 1, 'shares': 11051},
#      'net':    {'count': -10, 'shares': -13121499}},
#     ...
#   ],
#   'holders': [
#     {'manager': 'Warren Buffett', 'firm': 'Berkshire Hathaway',
#      'portfolio_pct': 22.60, 'activity': 'Reduce',
#      'activity_pct': 4.32, 'position_value': 61961735000},
#     ...
#   ]
# }

Activity types in quarterly_activity:

Type	Meaning
buy	New position opened
add	Existing position increased
reduce	Position decreased
sell	Full exit (0 shares remaining)
net	(buy + add) − (reduce + sell)

Holder fields:

Field	Description
manager	Investor name
firm	Fund/firm name
portfolio_pct	% of this investor's portfolio in this stock
activity	Most recent action: Buy, Add, Reduce, Sell, or None
activity_pct	Magnitude of the action (%)
position_value	Current position value ($)

.managers() → list[dict]

List all 82 superinvestors tracked by DataRoma.

si.managers()
# [{'name': 'Warren Buffett', 'firm': 'Berkshire Hathaway', 'code': 'BRK',
#   'portfolio_value': '$274 B', 'num_stocks': 42,
#   'top_holdings': ['AAPL', 'AXP', 'BAC', 'KO', 'CVX', ...]}, ...]

---

CLI

All functions are available from the command line.

# Consensus buys (quarterly, top 5)
superinvestor buys -n 5

# Consensus buys (6 months)
superinvestor buys --period 6m

# Consensus sells
superinvestor sells -n 10

# Stock deep dive
superinvestor stock AAPL

# Grand portfolio
superinvestor holdings -n 20

# All superinvestors
superinvestor managers

# JSON output (for piping to jq, scripts, etc.)
superinvestor buys -n 5 --json

---

How It Works

The Consensus Thesis

"If one investor buys a stock, it could mean anything. If ten legendary investors independently buy the same stock, pay attention."

This library is built around the idea that agreement among multiple independent, skilled investors is a stronger signal than any single investor's picks. SEC 13F filings (required for institutional managers with $100M+ AUM) reveal what these investors actually own — not what they say on TV.

Data Source

All data comes from DataRoma, which aggregates SEC 13F filings for 82 carefully curated "superinvestors" including:

• Warren Buffett (Berkshire Hathaway)
• Bill Ackman (Pershing Square)
• David Einhorn (Greenlight Capital)
• Seth Klarman (Baupost Group)
• Li Lu (Himalaya Capital)
• Terry Smith (Fundsmith)
• And 76 more...

Update Frequency

13F filings are submitted quarterly, within 45 days of quarter end. DataRoma processes them as they become available. Real-time data reflects the latest filings.

---

Disclaimer

This project is not affiliated with DataRoma or any of the investors tracked. Data is sourced from public SEC filings via DataRoma.

This tool is for educational and research purposes only. It is not financial advice. Past holdings do not predict future performance.

---

License

MIT