Order-Book Feed Engineering

7 algorithms in Market Data Engineering.

In this family#

1. Trade-and-Quote Event Normalization contract

   Normalises raw venue trade and quote messages into one canonical event shape. Every venue names, orders and timestamps its fields differently; this is the boundary where that stops being everyone else's problem.

   normalizeEvents(events)
2. Level-2 Snapshot-and-Delta Reconstruction contract

   Rebuilds an aggregated price-level book from a snapshot plus the deltas that follow it. Level 2 shows quantity per price, not individual orders — the distinction matters because a delta that is applied twice corrupts the book invisibly.

   reconstructL2(snapshot)
3. Level-3 Order-by-Order Reconstruction contract

   Rebuilds the book order by order rather than by price level. Level 3 preserves queue position, which is the only way to answer where in the queue an order actually sits — and therefore the only basis for a realistic fill model.

   reconstructL3(snapshot_orders)
4. Sequence-Gap Detection and Recovery contract

   Detects missing sequence numbers and decides whether the stream can continue or must resynchronise from a snapshot. A gap is not a nuisance — every message after it is applied to a book that is already wrong.

   recoverSequenceStream(arrivals)
5. Price-Level Quantity Aggregation contract

   Collapses individual orders into quantity per price level — the Level 3 to Level 2 projection, and the form most analytics actually consume.

   aggregatePriceLevels(orders)
6. Snapshot/Incremental-Feed Reconciliation contract

   Reconciles a periodic snapshot against the book built from incrementals. Any divergence means the incremental path has been silently wrong — possibly for hours — and this is the only routine that catches it.

   reconcileSnapshotIncrementals(snapshot)
7. Multi-Venue Best-Quote and Book Consolidation contract

   Merges books from several venues into one consolidated view and identifies the best bid and offer across them. Venue clocks differ, so a naive merge can produce a consolidated book that was never simultaneously true.

   consolidateVenues(quotes)

What they share#

Every topic here is a record-transform, so once you have called one the rest follow the same shape. Import paths differ only in the final segment:

import { normalizeEvents } from "fintech-algorithms/market-data-engineering/order-book-feed-engineering/trade-and-quote-event-normalization";
import { reconstructL2 } from "fintech-algorithms/market-data-engineering/order-book-feed-engineering/level-2-snapshot-and-delta-reconstruction";

Read them in the order above — the sequence is pedagogical, not alphabetical.

Where this sits#

Market Data Engineering collects 31 algorithms across 5 families. For the concept behind this family rather than the call signatures, see the concept guides.