A few personal projects I made from Jan 2022 to Sep 2022:
Gold Predictor => Time Series
- Aim: Predict the futre price of gold, this is a multivariant forecasting model , it uses other macro factors that affect gold e.g the strength of the dollar , to predict the price of gold. Forecasting for stock prices can be very difficult as there are many factors to consider and thus data to gather
Automatic Speech Recognition => Audio Classification
- Aim: Able to predict (from 10 words) the word spoken by anybody, it does this by processing the audio in to different forms ie. waveforms , spectograms , mel-spectrograms etc , this system can be used in the same way "Hey Siri" works, as an initiating word/phrase to activate other models, thus increasing computational efficiency
Object Detection
- Aim: This notebook demostrates the ability to create, train and utlise a deskop camera to create an facial detection model that can be configured for other use-cases
Spotify Sentiment Analysis => NLP
- Aim: Is able to categorise a music review from bad, avergae , good (0-1, 2-3, 4-5 stars) , if the user just left a comment with no rating.
Twitter Disaster Identification => NLP
- Aim: Is able to identify if a tweet is a disaster or not, and use the information from that to make a certain tweets trend more and quicker in order to inform more people of the disaster.
Traffic Sign Identification => Classification
- Aim: idenitfy traffic signs automatically, it can be used in self driving cars , or even normal cars to help drivers better drive (especially at night).
101Food => Classification
- Aim: Identify 101 classes of food just by showing the trained model a picture of the food.
international_airline_passengers_prediction => Basic time series
- Aim: Very basic, it aims to forecast the number of passengers at a given time to help airlines better understand customer seasonality trends which results in better planning and overal tracking of KPIs , the dataset has a seasonality trend making it easier to train the model on the data