pytorch tabular provides a unified framework for deep learning models on tabular data.
Open source
24 alternativesDisco
Disco’s discovery engine finds statistically validated patterns in tabular data, including feature interactions and subgroup effects.
About Disco
Disco is the Discovery Engine published by Leap Laboratories as an open-source repository. It searches tabular data for combinations of feature conditions that significantly shift a chosen target column, rather than starting from a user's hypothesis. Typical findings include feature interactions, subgroup effects and conditional relationships, expressed with specific thresholds instead of vague correlations.
Each pattern is validated on a hold-out set, carries FDR-corrected p-values and effect sizes, and is checked against academic literature and classified as novel or confirmatory, with citations. Results also link to an interactive web report. Analyses run through an API key and take a few minutes; the Python client polls automatically and logs queue position, current pipeline step and ETA. The project states it is not intended for summary statistics, visualisation, filtering or SQL queries.
Key features
- Finds combinations of feature conditions that shift a target column
- Reports FDR-corrected p-values and effect sizes per pattern
- Validates each pattern on a hold-out set
- Checks patterns against academic literature and classifies novelty
- Python SDK polls queued runs and logs progress and ETA
- Links results to an interactive web report of patterns
Best for: Analysts and data scientists with tabular datasets who want to surface validated interactions and subgroup effects without pre-specifying a hypothesis.
Good for
- Open-source repository with an installable Python package (discovery-engine-api)
- Patterns include hold-out validation, FDR-corrected p-values and support counts
- Each pattern comes with literature citations and a novel-versus-confirmatory classification
Watch out for
- An API key obtained through email verification is required to run analyses
- Explicitly not intended for summary statistics, visualisation, filtering or SQL queries
- Analyses run on a queue and take a few minutes rather than returning instantly
Disco at a glance
- Category
- AI & Machine Learning › AI Data & Analytics
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Open source
- Yes · Source code
- Language
- English
- Last checked
- Oct 6, 2026 · Online when checked
- Listed
- Oct 5, 2026
Disco alternatives
Tools and sites you could use instead, ranked by how closely they match.
PyTorch Frame provides a library for applying deep learning models to tabular data.
Open source
24 alternatives
This TabGAN repository reviews research on applying GANs to generate tabular data.
Open source
24 alternatives
Apache Spark is a unified engine for large-scale data processing and analytics.
Open source
24 alternatives
Adala automates data labeling through an autonomous agent framework for organizing labeled datasets.
Open source
24 alternatives
tsfresh automatically extracts relevant features from time series data.
Open source
24 alternativesPopular in AI Data & Analytics
Well-known listings from the same topic.
Dotgetdot.aiForvioforvio.comFluidAI Medicalfluidai.mdApache Zeppelinzeppelin.apache.orgSola Securitysola.securityQuadraticquadratichq.com
Disco FAQ
What does each returned pattern contain?
A description, the feature conditions and their thresholds, a hold-out-validated FDR-corrected p-value, effect direction and size, support count and percentage, a novelty type with an explanation, and literature citations.
What is Disco not meant for?
The project lists summary statistics, visualisation, filtering and SQL queries as out of scope, recommending pandas for those tasks.
How do you run a first analysis?
Install the discovery-engine-api package, obtain an API key by email verification or from the developers page, then call the engine's discover method with a file and target column.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.