Skip to content

First query

Analyst

Use this page for a linear first session against an existing local warehouse. OddsFox Pipeline ships software and local warehouse tooling, not a hosted dataset. For table selection, trust rules, and the full chooser, continue with Query the warehouse.

1. Open read-only

Prefer a read-only connection so analysis does not compete with a writer:

import duckdb

con = duckdb.connect("oddsfox.duckdb", read_only=True)

Or open the CLI:

duckdb oddsfox.duckdb

The default path is oddsfox.duckdb in the repository root. If .env sets DUCKDB_PATH, open that file instead. See Configuration for path precedence.

If you do not have a warehouse yet, ask an operator to complete Quickstart, then return here.

2. List public schemas

show schemas;

Prefer schemas ending in *_marts. Use *_observability for freshness and quality checks. Treat *_raw, *_ops, *_staging, and *_intermediate as internal surfaces.

3. Pick a starting mart

Goal Start here
Polymarket WC2026 hourly odds polymarket_wc2026_marts.polymarket_wc2026_market_hourly_odds
Current Kalshi stage prices kalshi_wc2026_marts.kalshi_wc2026_stage_markets

The full chooser lives in Query the warehouse.

4. Run one query

Polymarket golden mart (after a Polymarket full run):

rows = con.sql("""
    select
        event_slug,
        question,
        odds_hour_utc,
        close_odds,
        event_volume_usd_lifetime_reported
    from polymarket_wc2026_marts.polymarket_wc2026_market_hourly_odds
    where is_active
      and not is_closed
    order by event_slug, question, odds_hour_epoch desc
""").df()

Kalshi stage markets (after a Kalshi full run) — prefer is_actionable_live_market:

select
    canonical_team_name,
    stage_key,
    progression_outcome_label,
    progression_price,
    current_price_status,
    market_ticker
from kalshi_wc2026_marts.kalshi_wc2026_stage_markets
where is_actionable_live_market
order by canonical_team_name, stage_rank;

5. Trust check

Before treating prices as live, inspect observability:

select *
from polymarket_wc2026_observability.polymarket_wc2026_ingestion_run_observability
order by 1 desc
limit 20;

select *
from kalshi_wc2026_observability.kalshi_wc2026_data_quality
limit 50;

Historical closed and resolved rows are intentionally retained. Do not assume a row is live because it appears in a mart. Term shortcuts live in Analyst shortcuts.

6. Next steps

Goal Page
Copy-paste SQL and Python Query recipes
Grain, filters, and common mistakes Data dictionary
Full table chooser and trust rules Query the warehouse
Analyst map and joins Analysts

See also