Row2Vec: Learn Embeddings from Tabular Data

Row2Vec: Learn Embeddings from Tabular Data#

Welcome

Row2Vec is a Python library for easily generating low-dimensional vector embeddings from any tabular dataset. It uses deep learning and classical ML methods to create powerful, dense representations of your data.

What is Row2Vec?#

Row2Vec transforms tabular data into meaningful vector representations (embeddings) that capture the essential characteristics of your data. Instead of feeding raw data directly into models, you can create compressed, information-rich representations that models can easily process.

Key Features#

  • 🎯 Five Powerful Modes: Neural networks, PCA, t-SNE, UMAP, and target-based embeddings

  • 🧠 Intelligent Preprocessing: Automatic missing value imputation and feature encoding

  • 🚀 Simple API: One function - learn_embedding() - handles everything

  • 💾 Model Persistence: Save and load trained models for production use

  • 🔧 Production Ready: 92% test coverage, type safety, modern build system

Quick Example#

# Import complete suppression first
exec(open('suppress_minimal.py').read())

import pandas as pd
from row2vec import learn_embedding, generate_synthetic_data

# Generate sample data
df = generate_synthetic_data(num_records=100, seed=42)
print(f"Data shape: {df.shape}")
print(df.head(3))

✓ Enhanced minimal suppression active

Data shape: (100, 3)
  Country Product       Sales
0     USA       A  103.047171
1  Mexico       B  448.000795
2  Canada       B   80.489648

# Generate unsupervised embeddings
embeddings = learn_embedding(df, mode="unsupervised", embedding_dim=2, max_epochs=10, verbose=False)
print(f"\nEmbedding shape: {embeddings.shape}")
print("\nFirst 5 embeddings:")
print(embeddings.head())

Model: "functional"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ input_layer (InputLayer)        │ (None, 10)             │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense (Dense)                   │ (None, 128)            │         1,408 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout (Dropout)               │ (None, 128)            │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ embedding (Dense)               │ (None, 2)              │           258 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_1 (Dense)                 │ (None, 128)            │           384 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout_1 (Dropout)             │ (None, 128)            │             0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_2 (Dense)                 │ (None, 10)             │         1,290 │
└─────────────────────────────────┴────────────────────────┴───────────────┘

 Total params: 3,340 (13.05 KB)

 Trainable params: 3,340 (13.05 KB)

 Non-trainable params: 0 (0.00 B)

Embedding shape: (100, 2)

First 5 embeddings:
   embedding_0  embedding_1
0    -0.052516     0.156451
1     0.094666     0.168608
2    -0.419540    -0.065374
3    -0.074119     0.141072
4    -0.153844    -0.141039

Why Use Row2Vec?#

Compared to Manual Implementation#

Aspect

Row2Vec

Manual Neural Network

Lines of code

~5

~200+

Preprocessing

Automatic

Manual pipeline

Missing values

Handled

Manual imputation

Categorical encoding

Automatic

Manual encoding

Scaling

Built-in

Manual setup

Compared to Other Methods#

Method

Use Case

Row2Vec Advantage

PCA

Linear reduction

Also offers non-linear (neural) options

t-SNE

Visualization

Unified interface with preprocessing

UMAP

General reduction

Consistent API across all methods

Manual NN

Custom embeddings

Automatic preprocessing, simpler API

Installation#

pip install row2vec

Documentation Overview#

Next Steps#

Ready to get started? Head to the Installation guide or jump straight to the Quickstart tutorial.

Questions or Issues?

Contents