EDA AI Hub

EDA AI Hub is a playground for Exploratory Data Analysis (EDA), Machine Learning and Deep Learning projects (AI), using a common code base for multiple jupyter notebooks.

Projects

Sample projects included:

• Image Caption Generation using COCO dataset (Deep Learning)

  • Uses a subset of the COCO "Common Objects in Context" dataset for image caption generation
  • Features transfer learning, by using a pretrained ResNet152 on the ImageNet dataset
  • Compares prediction results using BLEU score and cosine similarity

• DARWIN (Diagnosis Alzheimer's With Handwriting) (Machine Learning)

  • Analyzes Alzheimer's patients handwriting patterns using the DARWIN dataset
  • Features in-depth overfitting analysis, i.e.: PCA vs NonPCA
  • Compares Logistic Regression and Support Vector Classifier results

• Adventure Works 2022 business analysis (Data Science)

  • Uses the Adventure Works 2022 dataset and provides sports equipment sales metrics and business information.

Note that these projects are extensive, so it's recommended to use an IDE's Outline view for navigating notebook sections.

Learning examples

Simpler learning examples using Kaggle datasts, include:

• GPU vs CPU analysis
• NASA Facilities dataset analysis
• World Cities statistics

Refer to the available jupyter notebooks (*.ipynb) for details.

Requirements

We need to install all the required python modules.
This can be done by executing make requirements in the base project folder.

cd $HOME/workspace/eda-hub  
make requirements