onlinerake 1.4.0

Streaming survey raking via SGD and MWU

onlinerake

Real-time survey weighting for streaming data.

The Problem

You're collecting survey responses or observational data one record at a time. Your sample doesn't match population demographics—too many young respondents, too few from certain regions. Traditional weighting methods (raking/IPF) require reprocessing the entire dataset whenever a new response arrives.

onlinerake updates weights incrementally as each observation streams in, keeping weighted margins aligned with population targets in real time.

When to Use This

• Online surveys where responses arrive continuously
• A/B tests that need demographic balance during collection
• Passive data collection (app usage, sensor data) requiring real-time calibration
• Any streaming scenario where batch reweighting is too slow or impractical

Quick Start

pip install onlinerake

from onlinerake import OnlineRakingSGD, Targets

# Define population targets (proportion with indicator = 1)
targets = Targets(
    female=0.51,      # 51% female in population
    college=0.32,     # 32% college educated
    age_65_plus=0.17  # 17% age 65+
)

# Create raker
raker = OnlineRakingSGD(targets, learning_rate=5.0)

# Process observations as they arrive
for response in survey_stream:
    raker.partial_fit(response)

    # Check current state anytime
    print(f"Weighted margins: {raker.margins}")
    print(f"Effective sample size: {raker.effective_sample_size:.0f}")

# Get final weights
weights = raker.weights[:raker.n_obs]

Which Algorithm?

Use Case	Algorithm	Learning Rate
Most cases	OnlineRakingSGD	5.0
Smoother weights, higher ESS	OnlineRakingSGD	2.0-5.0
IPF-like multiplicative updates	OnlineRakingMWU	0.5-1.0
Starting from unequal base weights	OnlineRakingMWU	0.5-1.0

Recommendation: Start with OnlineRakingSGD(targets, learning_rate=5.0). It converges faster, maintains higher effective sample size, and handles most scenarios well.

Performance

In simulation studies across linear drift, sudden shift, and oscillating bias scenarios:

Method	Margin Error Reduction	Effective Sample Size
SGD	72-80%	225-280 (of 300)
MWU	47-52%	175-276 (of 300)
Unweighted	baseline	300

SGD consistently outperforms MWU on margin accuracy while maintaining comparable effective sample sizes.

Features

Continuous Covariates (v1.3.0)

Target means instead of proportions:

targets = Targets(
    age=(42.0, "mean"),      # Target mean age = 42
    income=(55000, "mean"),  # Target mean income = $55,000
    female=0.51              # Binary: 51% female
)

Learning Rate Schedules

For theoretical convergence guarantees:

from onlinerake import OnlineRakingSGD, Targets, PolynomialDecayLR
from onlinerake.convergence import verify_robbins_monro

schedule = PolynomialDecayLR(initial_lr=10.0, power=0.6)
raker = OnlineRakingSGD(targets, learning_rate=schedule)

# Verify Robbins-Monro conditions (analytical for known schedules)
result = verify_robbins_monro(schedule)
print(result.condition_1_satisfied)  # True: Σ η_t = ∞
print(result.condition_2_satisfied)  # True: Σ η_t² < ∞

The verify_robbins_monro() function provides analytical verification for known schedule types with mathematical proofs.

Diagnostics

from onlinerake import check_target_feasibility, compute_design_effect

# Check if targets are achievable with your data
feasibility = check_target_feasibility(raker)
print(f"Feasible: {feasibility.is_feasible}")

# Measure weighting efficiency
deff = compute_design_effect(raker)
print(f"Design effect: {deff:.2f}")

Batch Comparison

Compare streaming results against traditional IPF:

from onlinerake import BatchIPF

batch_raker = BatchIPF(targets)
batch_raker.fit(all_observations)

print(f"Online loss: {online_raker.loss:.6f}")
print(f"Batch loss: {batch_raker.loss:.6f}")

API Reference

Core Classes

Targets(**features) - Define population margins

• Binary features: female=0.51 (proportion = 1)
• Continuous features: age=(42.0, "mean") (target mean)

OnlineRakingSGD(targets, learning_rate=5.0) - SGD-based streaming raker

• .partial_fit(obs) - Process one observation
• .margins - Current weighted margins (dict)
• .loss - Current squared-error loss
• .weights - Weight array (use [:raker.n_obs] to slice)
• .effective_sample_size - ESS accounting for weight variation
• .converged - Whether loss is below tolerance

OnlineRakingMWU(targets, learning_rate=1.0) - Multiplicative weights raker

• Same API as OnlineRakingSGD

Key Parameters

Parameter	Default	Description
learning_rate	5.0 (SGD), 1.0 (MWU)	Step size for updates
min_weight	0.1	Minimum allowed weight
max_weight	10.0	Maximum allowed weight
n_steps	3	Gradient steps per observation
convergence_tol	1e-6	Loss threshold for convergence

Installation

pip install onlinerake

Development install:

git clone https://github.com/finite-sample/onlinerake.git
cd onlinerake
pip install -e ".[docs]"

Testing

pytest tests/ -v

Examples

See examples/ for complete worked examples:

• real_survey_example.py - Basic survey weighting
• ab_test_calibration.py - Balancing treatment/control groups
• ad_targeting_calibration.py - Real-time ad delivery calibration
• recommendation_balancing.py - Content recommendation fairness

Interactive notebooks in docs/notebooks/:

• 01_getting_started.ipynb - Visual introduction
• 02_performance_comparison.ipynb - Algorithm benchmarking
• 03_advanced_diagnostics.ipynb - Convergence and diagnostics

Citation

If you use this package in research, please cite:

@software{onlinerake,
  author = {Sood, Gaurav},
  title = {onlinerake: Streaming Survey Raking},
  url = {https://github.com/finite-sample/onlinerake},
  version = {1.3.0},
  year = {2026}
}

License

MIT