L2 — Order Book

Multi-level order book depth, liquidity, and imbalance.

Order book analytics aggregated across 5, 10, 20, and 25 levels — capturing depth, imbalance, and liquidity dynamics at each granularity without storing terabytes of raw snapshots.

Browse catalog API documentation

Shown: the median perpetual’s book depth across the top 25 levels, as a share of its 20:00–20:45 average, per venue, 18:00–01:00 UTC on 10 Oct 2025; minutes under half of normal take the strong tone, and the lowest of all, OKX’s 8.5% at 22:06, is orange · built with l2_liquidity 1m

L2 — Order Book Datasets

L2 Order Book Imbalance

Prime

Multi-depth (5, 10, 20, 25 levels) order book imbalance, ratio, and averages.

Imbalance (5 levels)
Imbalance (20 levels)
Imbalance (25 levels)
Imbalance Ratio (5 levels)

L2 Order Book Liquidity

Prime

Total bid/ask depth aggregated over 5, 10, 20, 25 order book levels — instantaneous and interval-averaged.

Ask Aggregate (10 levels)
Ask Aggregate (20 levels)
Ask Aggregate (25 levels)
Bid Aggregate (10 levels)

Highlighted Metrics

imbalance_5

Imbalance (5 levels)

Imbalance across five levels extends top-of-book pressure into the near-touch order book.

It reveals whether directional support is just a quote-level artifact or is backed by additional nearby depth.

imbalance_ratio_25

Imbalance Ratio (25 levels)

The twenty-five-level imbalance ratio normalizes deep-book pressure over a broader slice of resting liquidity.

It is useful for understanding the full displayed environment that larger orders would have to push through.

bid_ask_ratio_25

Bid/Ask Ratio (25 levels)

The twenty-five-level ratio extends that comparison deep enough to capture broader displayed inventory.

It is especially relevant for larger execution and for understanding whether apparent pressure near the touch is reinforced deeper in the ladder.

ask_agg_25_avg

Avg Ask Aggregate (25 levels)

Average Ask Aggregate across twenty-five levels smooths the deeper sell-side book over the interval.

That makes it a regime measure for how much overhead liquidity the market was consistently showing, not just flashing moment to moment.

bid_agg_25_avg

Avg Bid Aggregate (25 levels)

Average Bid Aggregate across twenty-five levels provides the broad matching view of displayed support.

Stronger readings indicate a deeper cushion of resting liquidity that can absorb pressure before the book thins out materially.

Market Microstructure Context

Order book imbalance

Cont, Kukanov & Stoikov (2014) showed it clearly: order book imbalance predicts the direction of the next mid-price change, and the relationship is monotonic — stronger lean, larger expected move. The mechanism is straightforward. When one side is substantially thicker, the thinner side depletes first and the midpoint shifts. But the real insight is in the layers. L1 and L2 can tell contradictory stories, and the divergence between them is often more informative than either alone.

L1 imbalance — Best-bid versus best-ask quantity. A single snapshot is noisy; time-averaged imbalance over the interval is what actually predicts.
L2 imbalance — The same signal extended across 5, 10, 20, and 25 levels per side. Captures the full shape of supply and demand behind the frontier — L2 can be draining even while L1 looks stable.
L1-L2 divergence — Where the actionable signal lives. Thin L1 bids with deep L2 behind them is a market maker repositioning, not genuine weakness. Thick L1 bids with nothing behind them is a single concentrated quote — pull it and there is a gap.
Weighted midprice — Stoikov (2018) showed the quantity-weighted midprice — shifted proportional to size at each side — beats the simple midpoint as a predictor of the next transaction price.

Predicting Short-Term Crypto Returns with Market Microstructure

Interactive notebook

Use Cases

Deep liquidity analysis

Measure aggregate bid and ask depth at multiple levels to understand resilience and liquidity beyond the top-of-book.

Order book imbalance signals

Build predictive features from multi-level imbalance ratios that capture directional pressure more robustly than L1 alone.

Iceberg & hidden order detection

Detect large resting orders by monitoring depth anomalies and imbalance dynamics across book levels over time.

Market resilience measurement

Quantify how quickly liquidity replenishes after consumption by analyzing depth averages and their variability.

Cross-level depth feature engineering

Combine bid/ask depth at 5, 10, 20, 25 levels into composite signals for ML-based execution and alpha models.

Get started in minutes

curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | bash

github ↗

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Fetch the preview slice as Parquet into ./data — no API key needed
aperiodic flow --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --interval 5m \
  --timestamp exchange \
  --start-date 2025-05-01 \
  --end-date 2025-05-31 \
  --output-dir ./data && echo "Saved to $PWD/data"

Preview · binance-futures · perpetual-BTC-USDT:USDT · 1h · 2024-01-01 → 2024-01-31

df.head() · 5 × 37

#	time	taker_buy_volume	taker_sell_volume	taker_buy_count	taker_sell_count	volume_delta	volume_delta_notional	flow_entropy	flow_toxicity_score	taker_buy_sell_ratio	taker_buy_sell_count_ratio	taker_buy_sell_percentage	taker_buy_sell_count_percentage	taker_buy_small_order_volume	taker_buy_small_order_count	taker_buy_medium_order_volume	taker_buy_medium_order_count	taker_buy_large_order_volume	taker_buy_large_order_count	taker_sell_small_order_volume	taker_sell_small_order_count	taker_sell_medium_order_volume	taker_sell_medium_order_count	taker_sell_large_order_volume	taker_sell_large_order_count	taker_buy_small_order_percentage	taker_buy_medium_order_percentage	taker_buy_large_order_percentage	taker_sell_small_order_percentage	taker_sell_medium_order_percentage	taker_sell_large_order_percentage	taker_buy_small_order_count_percentage	taker_buy_medium_order_count_percentage	taker_buy_large_order_count_percentage	taker_sell_small_order_count_percentage	taker_sell_medium_order_count_percentage	taker_sell_large_order_count_percentage
1	2024-01-01 00:00:00	4687.97607421875	3771.5009765625	46500	41778	916.4749755859375	38871268	0.6872671842575073		1.2430001497268677	1.1130260229110718	0.5541685819625854	0.526745080947876	138008.609375	2684	9790435	26418	189105360	17398	117024.328125	1897	8532792	24171	151512720	15710	29.438846588134766	2088.4140625	40338.37890625	31.028581619262695	2262.439208984375	40173.0546875	0.057720430195331573	0.5681290030479431	0.3741505444049835	0.04540667310357094	0.578558087348938	0.37603524327278137
2	2024-01-01 01:00:00	4783.837890625	4259.5732421875	49386	40965	524.2650146484375	22391096	0.6914658546447754	-0.5	1.1230791807174683	1.2055656909942627	0.5289859771728516	0.5466015934944153	140995.4375	2838	10853161	28316	193186432	18232	77478.265625	1240	9094397	23605	172617616	16120	29.473289489746094	2268.714111328125	40383.1484375	18.189208984375	2135.048828125	40524.625	0.057465679943561554	0.5733608603477478	0.36917343735694885	0.030269742012023926	0.576223611831665	0.39350664615631104
3	2024-01-01 02:00:00	2141.259033203125	2511.80810546875	24485	28065	-370.54901123046875	-15779090	0.6899729371070862	0.9461790323257446	0.8524771332740784	0.8724389672279358	0.460182249546051	0.46593719720840454	81757.6171875	1594	5058392.5	13066	86080624	9825	42772.078125	685	6032806.5	16578	100924288	10802	38.182029724121094	2362.344970703125	40200.9375	17.02840232849121	2401.7783203125	40179.9375	0.06510107964277267	0.5336328148841858	0.40126606822013855	0.024407625198364258	0.5907001495361328	0.38489222526550293
4	2024-01-01 03:00:00	3501.132080078125	4618.748046875	36157	48188	-1117.615966796875	-47477344	0.683644711971283	-0.17025771737098694	0.7580261826515198	0.7503320574760437	0.43118026852607727	0.42867982387542725	109771.1796875	2160	7630267.5	20143	140732512	13854	78518.9765625	1253	10414373	29296	185457008	17639	31.353052139282227	2179.37158203125	40196.2890625	17.00005531311035	2254.80419921875	40153.08984375	0.05973947048187256	0.5570982098579407	0.38316231966018677	0.026002325117588043	0.6079521775245667	0.3660454750061035
5	2024-01-01 04:00:00	3155.431884765625	3201.10400390625	36640	31386	-45.672000885009766	-1877300.75	0.6931213736534119	0.026555629447102547	0.98573237657547	1.167399525642395	0.4964074492454529	0.5386176109313965	150556.921875	3042	7435109	20135	126048920	13463	59707.23828125	946	6590907	18784	128861272	11656	47.71356964111328	2356.288818359375	39946.64453125	18.652076721191406	2058.947998046875	40255.2578125	0.08302401751279831	0.5495360493659973	0.36743995547294617	0.030140826478600502	0.5984833836555481	0.37137576937675476

Ready to start building?

Get access to our full catalog of market microstructure data.

Get started free

Shown: minute lows of the 465 Binance perpetuals, 20:40–22:00 UTC on 10 Oct 2025, with BTC’s path traced · built with ohlcv 1m

L2 — Order Book

Multi-level order book depth, liquidity, and imbalance.

Order book analytics aggregated across 5, 10, 20, and 25 levels — capturing depth, imbalance, and liquidity dynamics at each granularity without storing terabytes of raw snapshots.

Browse catalog API documentation

L2 — Order Book Datasets

L2 Order Book Imbalance

Prime

Multi-depth (5, 10, 20, 25 levels) order book imbalance, ratio, and averages.

Imbalance (5 levels)
Imbalance (20 levels)
Imbalance (25 levels)
Imbalance Ratio (5 levels)

L2 Order Book Liquidity

Prime

Total bid/ask depth aggregated over 5, 10, 20, 25 order book levels — instantaneous and interval-averaged.

Ask Aggregate (10 levels)
Ask Aggregate (20 levels)
Ask Aggregate (25 levels)
Bid Aggregate (10 levels)

Highlighted Metrics

imbalance_5

Imbalance (5 levels)

Imbalance across five levels extends top-of-book pressure into the near-touch order book.

It reveals whether directional support is just a quote-level artifact or is backed by additional nearby depth.

imbalance_ratio_25

Imbalance Ratio (25 levels)

The twenty-five-level imbalance ratio normalizes deep-book pressure over a broader slice of resting liquidity.

It is useful for understanding the full displayed environment that larger orders would have to push through.

bid_ask_ratio_25

Bid/Ask Ratio (25 levels)

The twenty-five-level ratio extends that comparison deep enough to capture broader displayed inventory.

It is especially relevant for larger execution and for understanding whether apparent pressure near the touch is reinforced deeper in the ladder.

ask_agg_25_avg

Avg Ask Aggregate (25 levels)

Average Ask Aggregate across twenty-five levels smooths the deeper sell-side book over the interval.

That makes it a regime measure for how much overhead liquidity the market was consistently showing, not just flashing moment to moment.

bid_agg_25_avg

Avg Bid Aggregate (25 levels)

Average Bid Aggregate across twenty-five levels provides the broad matching view of displayed support.

Stronger readings indicate a deeper cushion of resting liquidity that can absorb pressure before the book thins out materially.

Market Microstructure Context

Order book imbalance

L1 imbalance — Best-bid versus best-ask quantity. A single snapshot is noisy; time-averaged imbalance over the interval is what actually predicts.
L2 imbalance — The same signal extended across 5, 10, 20, and 25 levels per side. Captures the full shape of supply and demand behind the frontier — L2 can be draining even while L1 looks stable.
L1-L2 divergence — Where the actionable signal lives. Thin L1 bids with deep L2 behind them is a market maker repositioning, not genuine weakness. Thick L1 bids with nothing behind them is a single concentrated quote — pull it and there is a gap.
Weighted midprice — Stoikov (2018) showed the quantity-weighted midprice — shifted proportional to size at each side — beats the simple midpoint as a predictor of the next transaction price.

Predicting Short-Term Crypto Returns with Market Microstructure

Interactive notebook

Use Cases

Deep liquidity analysis

Measure aggregate bid and ask depth at multiple levels to understand resilience and liquidity beyond the top-of-book.

Order book imbalance signals

Build predictive features from multi-level imbalance ratios that capture directional pressure more robustly than L1 alone.

Iceberg & hidden order detection

Detect large resting orders by monitoring depth anomalies and imbalance dynamics across book levels over time.

Market resilience measurement

Quantify how quickly liquidity replenishes after consumption by analyzing depth averages and their variability.

Cross-level depth feature engineering

Combine bid/ask depth at 5, 10, 20, 25 levels into composite signals for ML-based execution and alpha models.

Get started in minutes

curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | bash

github ↗

terminal

# Install the Aperiodic CLI into ~/.local/bin
mkdir -p "$HOME/.local/bin"
curl -fsSL https://raw.githubusercontent.com/aperiodic-io/cli/main/install.sh | INSTALL_DIR="$HOME/.local/bin" bash
export PATH="$HOME/.local/bin:$PATH"

# Fetch the preview slice as Parquet into ./data — no API key needed
aperiodic flow --preview \
  --exchange binance-futures \
  --symbol perpetual-BTC-USDT:USDT \
  --interval 5m \
  --timestamp exchange \
  --start-date 2025-05-01 \
  --end-date 2025-05-31 \
  --output-dir ./data && echo "Saved to $PWD/data"

Preview · binance-futures · perpetual-BTC-USDT:USDT · 1h · 2024-01-01 → 2024-01-31

df.head() · 5 × 37

#	time	taker_buy_volume	taker_sell_volume	taker_buy_count	taker_sell_count	volume_delta	volume_delta_notional	flow_entropy	flow_toxicity_score	taker_buy_sell_ratio	taker_buy_sell_count_ratio	taker_buy_sell_percentage	taker_buy_sell_count_percentage	taker_buy_small_order_volume	taker_buy_small_order_count	taker_buy_medium_order_volume	taker_buy_medium_order_count	taker_buy_large_order_volume	taker_buy_large_order_count	taker_sell_small_order_volume	taker_sell_small_order_count	taker_sell_medium_order_volume	taker_sell_medium_order_count	taker_sell_large_order_volume	taker_sell_large_order_count	taker_buy_small_order_percentage	taker_buy_medium_order_percentage	taker_buy_large_order_percentage	taker_sell_small_order_percentage	taker_sell_medium_order_percentage	taker_sell_large_order_percentage	taker_buy_small_order_count_percentage	taker_buy_medium_order_count_percentage	taker_buy_large_order_count_percentage	taker_sell_small_order_count_percentage	taker_sell_medium_order_count_percentage	taker_sell_large_order_count_percentage
1	2024-01-01 00:00:00	4687.97607421875	3771.5009765625	46500	41778	916.4749755859375	38871268	0.6872671842575073		1.2430001497268677	1.1130260229110718	0.5541685819625854	0.526745080947876	138008.609375	2684	9790435	26418	189105360	17398	117024.328125	1897	8532792	24171	151512720	15710	29.438846588134766	2088.4140625	40338.37890625	31.028581619262695	2262.439208984375	40173.0546875	0.057720430195331573	0.5681290030479431	0.3741505444049835	0.04540667310357094	0.578558087348938	0.37603524327278137
2	2024-01-01 01:00:00	4783.837890625	4259.5732421875	49386	40965	524.2650146484375	22391096	0.6914658546447754	-0.5	1.1230791807174683	1.2055656909942627	0.5289859771728516	0.5466015934944153	140995.4375	2838	10853161	28316	193186432	18232	77478.265625	1240	9094397	23605	172617616	16120	29.473289489746094	2268.714111328125	40383.1484375	18.189208984375	2135.048828125	40524.625	0.057465679943561554	0.5733608603477478	0.36917343735694885	0.030269742012023926	0.576223611831665	0.39350664615631104
3	2024-01-01 02:00:00	2141.259033203125	2511.80810546875	24485	28065	-370.54901123046875	-15779090	0.6899729371070862	0.9461790323257446	0.8524771332740784	0.8724389672279358	0.460182249546051	0.46593719720840454	81757.6171875	1594	5058392.5	13066	86080624	9825	42772.078125	685	6032806.5	16578	100924288	10802	38.182029724121094	2362.344970703125	40200.9375	17.02840232849121	2401.7783203125	40179.9375	0.06510107964277267	0.5336328148841858	0.40126606822013855	0.024407625198364258	0.5907001495361328	0.38489222526550293
4	2024-01-01 03:00:00	3501.132080078125	4618.748046875	36157	48188	-1117.615966796875	-47477344	0.683644711971283	-0.17025771737098694	0.7580261826515198	0.7503320574760437	0.43118026852607727	0.42867982387542725	109771.1796875	2160	7630267.5	20143	140732512	13854	78518.9765625	1253	10414373	29296	185457008	17639	31.353052139282227	2179.37158203125	40196.2890625	17.00005531311035	2254.80419921875	40153.08984375	0.05973947048187256	0.5570982098579407	0.38316231966018677	0.026002325117588043	0.6079521775245667	0.3660454750061035
5	2024-01-01 04:00:00	3155.431884765625	3201.10400390625	36640	31386	-45.672000885009766	-1877300.75	0.6931213736534119	0.026555629447102547	0.98573237657547	1.167399525642395	0.4964074492454529	0.5386176109313965	150556.921875	3042	7435109	20135	126048920	13463	59707.23828125	946	6590907	18784	128861272	11656	47.71356964111328	2356.288818359375	39946.64453125	18.652076721191406	2058.947998046875	40255.2578125	0.08302401751279831	0.5495360493659973	0.36743995547294617	0.030140826478600502	0.5984833836555481	0.37137576937675476

Ready to start building?

Get access to our full catalog of market microstructure data.

Get started free

Shown: minute lows of the 465 Binance perpetuals, 20:40–22:00 UTC on 10 Oct 2025, with BTC’s path traced · built with ohlcv 1m