10 Kaggle Datasets For Learning Python And Data Science

Photo by Joshua Sortino on Unsplash

I changed my career from accounting to data science without ever using Kaggle.

This was stupid.

Don’t make my mistake — use these 10 awesome datasets (including thoughts on what you can do with them).

1 — Amazon Reviews

• Calculate basic product analytics
• Use clustering algorithms to group products
• Endless NLP use cases: sentiment analysis, keyword extraction, summarization

2 — IBM Credit Card Fraud

• Develop framework for imbalanced data
• Build supervised ML model to predict fraud
• Use clustering algorithms to group consumers

3 — Soccer

• Create dashboard of advanced analytics
• Build a supervised ML model to predict outcomes
• Use clustering algorithms to group players

4 — World Food Facts

• Create dashboard of advanced analytics
• Build a time-series model to forecast prices
• Use clustering algorithms to group products

5 — Spotify Song Lyrics

• Create unique analytics using audio features
• Build neural networks to predict hits
• Use clustering algorithms to create sub-genres

6 — AMEX Default Prediction

• Build pipeline to handle large data
• Design sampling approach for imbalanced data
• Build supervised ML model to predict fraud

Use what you learned on IBM’s dataset and apply it to big data!

7 — Home Prices

• Calculate suite of analytics
• Create geospatial dashboard
• Build supervised ML model to predict prices

8 — FourSquare Location Matching

• Manipulate unstructured geospatial data
• Build geospatial analytics
• Use advanced clustering to group locations

9 — H&M Fashion

• Build pipeline to handle large data
• Calculate advanced analytics
• Create recommendation algorithm

10 — Crypto Market

• Calculate advanced analytics
• Build time-series forecasting model
• Build supervised ML model to predict prices