LikeFolio Data · No registration needed

Sample data

Real extracts, in the exact shape the API and bulk drops use, cut with the trial rules: rows end at 2026-04-09 (today minus six months). Entities, divisions and trends are named; only Google-search brand names are withheld. No registration needed.

Download the bundle (zip)

File	What it is	Rows / size
`coverage.csv`	The full coverage list: every entity with its `entity_id` (the bare ticker for single-division companies, `TICKER#division` otherwise), ticker, company, division, `is_primary` and ownership window.	659 entities
`company_demand_sample.csv`	All 16 metrics, every day since 2021-01-01, for four high-correlation companies — Z (Zillow), WING (Wingstop), CELH (Celsius) and COIN (Coinbase). Tidy-long with `kdate`, `raw`, `corrected` (whole numbers) and `flag`.	~114k rows
`trends_sample.csv`	Four high-correlation trends — buying crypto (tracks COIN), ordering delivery, taking a trip and home remodeling — daily since 2016-09-01, with corrections and flags.	~14k rows
`trend_catalog.csv`	All 141 trends: `trend_id`, `name`, `group`, `mapped_tickers`, `mapped_datasets`.	141 rows
`trend_ticker_map.csv`	The full trend → ticker map with a signed `weight` (positive = aligned, negative = inverse). Columns: `trend_id`, `name`, `ticker`, `weight`, `added_at`, `removed_at`. Some mapped tickers aren't yet in company-demand coverage — they're coverage targets (company data coming soon) and already carry trend data.	~1,450 pairs
`trend_dataset_map.csv`	The trend → public-dataset map: the Census / PCE / Fed series each trend is validated against.

Load it

import pandas as pd
d  = pd.read_csv("company_demand_sample.csv", parse_dates=["date"])
u  = pd.read_csv("coverage.csv")

# ticker-level demand = sum of core entities
tk = (d[d.metric.eq("mentions")]
        .groupby(["ticker", "date"]).corrected.sum()
        .unstack(0))

# what the cleaning touched
d[d.flag.ne("") & d.flag.notna()].groupby(["ticker", "flag"]).size()

# COIN purchase-intent vs the "buying crypto" trend — same demand, two sources
t = pd.read_csv("trends_sample.csv", parse_dates=["date"])
coin = tk["COIN"].resample("ME").sum()
crypto = t[t.name.eq("buying crypto")].set_index("date").corrected.resample("ME").sum()
coin.corr(crypto)   # ~0.75

What to check in the first hour

Join. Every entity_id in the demand file exists in coverage.csv. Roll up to ticker by summing entities.
Demand recovery (COIN). COIN purchase intent more than quadrupled from 2023 to 2025 (≈113k→476k), and mentions peaked in October 2025 — the signal rode the crypto recovery.
20212026
Housing cycle (Z). Zillow purchase intent fell about a third from 2021 to 2023 as mortgage rates spiked, then recovered through 2025 — the series tracks the housing-demand cycle.
20212026
A named demand event (WING). WING's Aug 30 – Sep 8 2022 spike is its nationwide chicken-sandwich launch — our reviewers flagged it demand_event and kept it in corrected (real demand, not marketing); shares ran roughly +32% in the two weeks after.
20212026
Demand led the tape (CELH). Celsius mentions tripled from 2021 to 2023 (≈31k→103k) as the energy drink went viral — peaking in July 2023, about eight months before the stock topped near $96 (from ~$20). When the buzz cooled through 2025, the shares did too.
20212026
The stay-home renovation boom (a trend). The home remodeling trend nearly doubled in spring 2020 (≈45k→88k/mo) and stayed elevated all year — the wave behind the home-improvement giants' multi-year runs off their March-2020 lows (Home Depot roughly 3×, Lowe's more than 4×).
20162026

When you want more tickers, more trends, or dates past the cutoff, a trial key unlocks the whole embargoed coverage through the API.