Not supported for mobile device.

Please use the website on a desktop or larger screen.

ML Hyper-Trainer

gamified machine learning

6 Challenges

MVP

MNIST Digits

🔢

· Target: 85% accuracy

Data Preview

Features

Scaling

Train

Dataset Overview

60,000 rows

20

TOTAL COLUMNS

19

NUMERIC FEATURES

2

MISSING COLUMNS

4

OUTLIER COLUMNS

Data Quality Issues Detected

Empty Column 1: 100% missing values

Empty Column 2: 98% missing values

Aspect Ratio: contains outliers

Intersections: contains outliers

Top-Left Corner: contains outliers

Bottom-Right Corner: contains outliers

ColumnTypeSample ValuesDistributionMissingOutliersImportance

Label

TARGET

Target digit (0–9)

Target
504

None

No

Center Pixels

Pixel region: center (high info)

Numeric
0255128

μ=87 σ=110

None

No

92%

Edge Pixels

Pixel region: edges (low info, mostly 0)

Numeric
000

μ=4.2 σ=22

None

No

21%

Mid Pixels

Pixel region: mid-ring

Numeric
1218064

μ=52 σ=87

None

No

68%

Mean Intensity

Average pixel intensity per image

Numeric
336148

μ=33 σ=28

None

No

74%

X-Symmetry

Left-right symmetry score

Numeric
0.80.90.4

μ=0.6 σ=0.2

None

No

45%

Y-Symmetry

Top-bottom symmetry score

Numeric
0.30.80.5

μ=0.4 σ=0.25

None

No

38%

Aspect Ratio

Bounding box width/height

Numeric
0.50.20.8

μ=0.5 σ=0.15

None

Yes

62%

Pixel Density

Ratio of non-zero pixels

Numeric
0.220.150.35

μ=0.25 σ=0.08

None

No

55%

Loop Count

Number of closed loops (0, 1, 2)

Numeric
010

μ=0.6 σ=0.6

None

No

85%

Intersections

Number of crossing points

Numeric
001

μ=0.3 σ=0.5

None

Yes

58%

Stroke Length

Total path length of digit

Numeric
12080150

μ=110 σ=35

None

No

41%

Mass Center X

X coordinate of pixel mass

Numeric
14.213.815.1

μ=14 σ=1.5

None

No

22%

Mass Center Y

Y coordinate of pixel mass

Numeric
14.514.113.8

μ=14.2 σ=1.8

None

No

19%

X Variance

Spread of pixels along X axis

Numeric
22.412.128.5

μ=20 σ=6

None

No

34%

Y Variance

Spread of pixels along Y axis

Numeric
35.242.128.4

μ=35 σ=5

None

No

31%

Empty Column 1

Completely missing feature

Numeric
?

μ=0 σ=0

100%

No

0%

Empty Column 2

Mostly missing feature

Numeric
?12

μ=12 σ=2

98%

No

0%

Top-Left Corner

Sum of top-left pixels

Numeric
025

μ=1.2 σ=8

None

Yes

5%

Bottom-Right Corner

Sum of bottom-right pixels

Numeric
045

μ=2.1 σ=12

None

Yes

3%

💡 Review the data carefully — understanding your features helps you make better preprocessing choices.

── PIPELINE SCORE ────

C

71/100

Accuracy modifier: ×1.05

Features

100

Scaling

65

Outliers

30

Architect

75

Remove low-importance features (<25%) to reduce noise.

Some features are highly skewed — try Log or Sqrt normalization.

You have outlier columns — consider clipping or imputing them.

Step 1 of 3

Score: 71/100