Chart a market's price

GET /v1/markets/{venue}/{id}/quotes returns asks recorded each time they changed, oldest first. Where the window reaches back before recording began, earlier points are hourly or daily closes with price_usd instead of asks. A step chart shows both honestly: each price holds until the next.

Python

import os
from datetime import datetime

import matplotlib.pyplot as plt
import requests

API = "https://api.routeur.app"
venue, market_id = "kalshi", os.environ["MARKET_ID"]

res = requests.get(f"{API}/v1/markets/{venue}/{market_id}/quotes", params={"days": 90},
                   headers={"X-API-Key": os.environ["ROUTEUR_API_KEY"]}, timeout=30)
res.raise_for_status()
quotes = res.json()["data"]

recorded, archived = [], []
for q in quotes:
    at = datetime.fromisoformat(q["observed_at"].replace("Z", "+00:00"))
    if q.get("yes_ask_usd"):
        recorded.append((at, float(q["yes_ask_usd"])))
    elif q.get("price_usd"):
        archived.append((at, float(q["price_usd"])))

fig, ax = plt.subplots(figsize=(10, 4))
if archived:
    ax.step(*zip(*archived), where="post", linestyle="--", label="Closing price (history)")
if recorded:
    ax.step(*zip(*recorded), where="post", label="Yes ask (recorded)")
ax.set_ylim(0, 1)
ax.set_ylabel("Price per $1 contract")
ax.set_title(market_id)
ax.legend()
plt.show()

Notes

At most 5,000 points per request. For long, busy windows, ask for fewer days.
The first point is the last change before the window, so the line starts at the left edge.
Asks, not trades. For what actually traded, read GET /v1/markets/{venue}/{id}/trades.

​Notes

Notes