Normally, the place rating system on yelp and google map calculate the overall rating score based on the simple average. It may not an effective guidance for users especially for choosing a restaurant. We focus on looking forward a personalized rating algorithm by considering a user's personality from various aspects. From our perspective, our personalized rating score makes real sense.

What it does

We calculate the restaurant rating based on 9 kinds of categories such as age, interest, income, etc.

How I built it

The method for calculation is a modified KNN algorithm we designed. Though knn is original a classifier tool, we adopt a way to achieve quantization of these categories. We assign tags based on different life aspects. Instead of determining classifications, we quantify each tag to estimate its influence to the total rate. For each input dataset, we find its Top-K closest training points and take their average as our recommend rate.

Challenges I ran into

Data Analysis and process.

Accomplishments that I'm proud of

Our idea has very practical meaning for people. Everybody has ever met the situation that they believed the rating from website but they really did not like the taste.

What's next for Personalized Restaurant Rating

Try to present it well.


The project website is for the presentation purpose on the localhost.

Share this project: