t-SNE for visualization in ML Python - Model Pipeline Trace

Model Pipeline - t-SNE for visualization

t-SNE is a tool that helps us see complex data by turning many features into just two or three, so we can plot and understand it better.

Data Flow - 3 Stages

1Input Data

1000 rows x 50 columns→Raw high-dimensional data with 50 features per sample→1000 rows x 50 columns

Sample 1: [0.5, 1.2, 3.3, ..., 0.7]

↓

2Preprocessing

1000 rows x 50 columns→Normalize each feature to have zero mean and unit variance→1000 rows x 50 columns

Sample 1 normalized: [-0.1, 0.3, 1.0, ..., -0.2]

↓

3t-SNE Embedding

1000 rows x 50 columns→Compute pairwise similarities and map data to 2D space preserving local structure→1000 rows x 2 columns

Sample 1 embedded: [12.3, -5.6]

Training Trace - Epoch by Epoch

Loss
1.2 |*       
1.0 | *      
0.8 |  **    
0.6 |   **   
0.4 |    **  
0.2 |     ** 
0.0 +--------
     1 5 10 20 Epochs

Epoch	Loss ↓	Accuracy ↑	Observation
1	1.2	N/A	Initial embedding with high loss, points scattered randomly
5	0.8	N/A	Clusters start to form, loss decreases as local similarities improve
10	0.5	N/A	Clearer clusters, loss steadily decreasing
20	0.3	N/A	Stable embedding, loss converges, clusters well separated

Prediction Trace - 3 Layers

Layer 1: Input Sample

Layer 2: Normalization

Layer 3: t-SNE Mapping

Model Quiz - 3 Questions

What does t-SNE mainly help us do with data?

A Increase the number of features

B See high-dimensional data in 2D or 3D

C Train a classification model

D Clean missing data

Key Insight

t-SNE transforms complex, high-dimensional data into a simple 2D map that reveals hidden groupings, making it easier to understand patterns visually.