Inspiration

We wanted to build something cool with Machine Learning and D3.js. Using a Kaggle Dataset for this purpose seemed the best way to go.

What it does

We try to tell how movies have evolved over the past 100 years. We also implemented a Machine Learning module that predicts movie revenues based on Director, Plot, Actor and Budget.

How we built it

The backend was written in Python, interfaced using Flask with the JS front end. We cleaned our dataset using Pandas. Built the JS visualisations with d3.js.

Challenges we ran into

It was a task identifying the correct model to use. Moreover, understanding Pandas and its neat tricks was another task. Additionally, d3.js had a complicated chaining syntax.

Accomplishments that we're proud of

Setting up Flask. Learning Javascript and D3.js. Data Analytics. Implementing machine learning models.

What we learned

Don't give up. Team work. One Hot Encoding and Dimensionality Curse. Flask, Numpy, Pandas.

What's next for AIMLDB

Using more sophisticated techniques and feature selection to predict better revenues.

We also got ourselves a domain - aimldb.com, but domain.com let us down. This is a wordplay on using AI/ML techniques to predict revenue from the iMDB database.

Share this project:
×

Updates