The Global Aggie City Recommender

• 

  Recommended Worldwide

• 

  Bar chart of ratings

• 

  Recommended America

• 

  Map America

• 

  All cities

• 

  Map Worldwide

Inspiration

Aggies, now more than ever, need tools to search cities they would prefer to live in. I thought it would be a great idea to see what I could come up with my knowledge of app building and machine learning

What it does

Based on how much importance you think should be given to various factors such as purchasing power, access to healthcare, pollution, crime, and overall quality of life, the app recommends cities that you should consider moving to.

How I built it

The app was built using Streamlit. The data wrangling and data analysis was done in Google Colab.

Challenges I ran into

Merging the two datasets- Movehub rating and Housing prices was a big challenge. Many cities in the US have names similar to ones in Europe, so filtering them out was tedious. I also learnt what can be put in streamlit cache and what cannot be. For example, you can't cache ML models, in my case this was the KMeans clustering model. So it is little expensive to train the model each time but thankfully, the dataset was not very big.

Accomplishments that I'm proud of

A nice app that gives personalized recommendations.

What I learned

Using streamlit to make nice front ends Using scikit-learn and pandas for processing datasets.

What's next for The Global Aggie City Recommender

I found a really cool dataset on the location of Chipotle restaurants on Kaggle. I would like to integrate this into my current data. I could potentially let the user decide how much Chipotle matters to them! I could also let the user decide which cities they want to live close to, then recommend based on given location.

Built With

• python
• scikit-learn
• streamlit