LunarCrush
Prepared for Brevan Howard
This data room is access-protected. Enter the access code from the email.
LunarCrush
For Brevan Howard
Confidential
Reddit Sentiment · Equities

Reddit-only sentiment. Full depth, per ticker, per day.

Network-isolated cut of the LunarCrush social intelligence feed for Brevan Howard: daily Reddit-only post volume, engagement, unique contributors, and classified sentiment (positive / net / negative) for five equities, plus a post-level sample with per-post sentiment scores and source links. Every number here is the full Reddit count for that day — not a sampled subset. Query everything in your browser below.

Tickers
5
NVDA · TSLA · AAPL · GME · AMD — any ticker or keyword can be cut identically
Window
90 d
daily buckets · hourly + multi-year available in production
Network
Reddit
isolated at source · same breakout for X, TikTok, YouTube, IG, news
Sentiment
3-way
positive / net / negative counts daily + 1–5 score on every post
Get the data

Download the files

Two datasets, each in Parquet (pandas / polars / DuckDB — loads in seconds) and TSV (opens directly in Excel). Schemas below; full methodology in the README.

Daily Reddit series — 5 tickers × 90 days

One row per ticker per day. Full network-isolated Reddit counts: how many posts were active and created, total engagement, unique accounts participating, and the day's sentiment classified three ways. This is the backtest-ready table.

445 rows · 10 columns · Mar 12 → Jun 9, 2026 · Parquet 12 KB · TSV 21 KB
date · symbol · reddit_posts_active posts receiving engagement that day · reddit_posts_created new posts · reddit_interactions upvotes + comments + shares · reddit_contributors_active unique accounts engaging · reddit_contributors_created unique accounts posting · reddit_sentiment_positive count of positive posts · reddit_sentiment net daily score · reddit_sentiment_negative count of negative posts
⬇ Parquet ⬇ TSV (Excel)

Post-level sample — sentiment on every post

The granularity behind the daily series: top Reddit posts per ticker per day with an individual sentiment score on each post (1–5 scale, 3 = neutral), engagement counts, and a link back to the original Reddit thread for spot-checking.

876 rows · 7 columns · 90-day window · Parquet 31 KB · TSV 121 KB
date · symbol · post_created_utc exact timestamp · post_sentiment 1–5 per-post score · interactions engagement on the post · title post title · link original Reddit URL
⬇ Parquet ⬇ TSV (Excel)

📄 README — schemas, methodology, and production-delivery options (hourly buckets, multi-year history, any ticker or keyword universe, all networks).

Daily series

Reddit activity by ticker

Last 14 days shown per ticker. The full 90-day series for all five tickers is in the daily table below and in the Parquet download.

DatePosts activePosts createdInteractionsContributorsSent +Sent netSent −
Loading…
SQL Playground

Query the data yourself

DuckDB running in your browser. Reads Parquet directly from object storage — no server, no auth, no data leaves your machine. Tables: daily (per-ticker daily Reddit series) and posts (post-level sample with per-post sentiment).