Here are projects I’ve worked on in the past.
Projects
Donor Prediction Challenge(click here to explore)
Github repo
Tasks:
Identify, based on historical data, the likely donors amongst a list of prospects. Extract insights on the most profitable demographics to target Challenges:
Messy and imbalanced dataset Tricks:
Data Cleaning Log transformation Undersampling Tools:
Python(Numpy, Pandas, Seaborn)
Projects
Electricity prediction- Time Series Forecasting Project (click here to explore)
Github repo
Tasks:
Forecasted energy consumption over a given period Created a dashboard for customers to view their historical energy usage Challenges:
Was my first time working with a time series dataset. Took some time to figure it out Tricks:
Used location clusters and Acorn classification to engineer new features Tools:
Python(Numpy, Pandas, Seaborn) SARIMA models Tensorflow(Keras backend) for LSTM Voila (dashboard for visualization)
Projects
Loan Prediction (click here to explore)
Github repo
Tasks:
Create classification framework to categorize customers Predict customers’ loan performance Challenges:
Imbalanced dataset with outliers Tricks:
Outlier removal Undersampling Tools:
Python(Numpy, Pandas, Seaborn, Matplotlib) R