A project from Training Data Ltd., a major online data science training provider, I tackled a challenge related to dataset size and model training times. One of their customer datasets was so large that models took days to generate predictions. My objective was to ensure this dataset was stored efficiently, allowing models to run on a reasonable timeframe without compromising data size.
Training Data Ltd. tasked me with cleaning up this massive customer dataset. The goal was to develop a proof-of-concept for a much more efficient storage solution. This solution would eventually be used to predict whether students were looking for a new job, allowing them to be directed to prospective recruiters.
I was provided with customer_train.csv, a subset of the entire dataset. This anonymized data contained student information and whether they were searching for a new job during their training.