Home

Welcome

In this repo we will explore the Yelp data-set from Kaggle a little bit.

Data Set

• excerpt of Yelp's businesses, reviews and user data
• for 11 metropolitan areas across 4 countries
• 5,200,000 user reviews on 174,000 businesses

Potential Analysis

Market / Businesses

1. Can we identify "trending" areas / neighbourhoods?
   • hip areas like Neukoeln, Kreuzberg etc.
2. Are the clusters of areas homogeneous in e.g. price level / categories / styles?
   • think of posh (Friedrichstrasse) vs. trashy (Warschauerstr.)
3. Would a retailer's store fit more into certain areas?
   • given a list of retailer venues, can we predict checkins (or sales even?)
4. Are there businesses that synergize?

Users

5. Do user affect / influence their friends (to test / visit venues)?
   • can we identify inluencers
6. Can we classify users into archetypes?
   • e.g. foodies, hipsters, fashionistas, parents, yuppies, etc.
7. Can we recommend businesses / venues to users?
   • predict user rating for biz
8. Can we predict whether / which businesses a user will visit next?
   • base on behaviour, preference, friends etc.