A list of projects I have done professionally and as time pass
- ACO observations- have experience working with these signals as time series before and only. Here raw data is aggregation VIN (vehicle identification number) and I have taken max, min and average statistics of ACO obtained from Vehicle over a period of time (see synthetic data for more details). performed EDA on this data - analyzing for road type on which high ACOs are observed, VIN count in high ACO zone (90th percentile)
- A bunch of forecasting time series projects - Understanding time series in any industry to analyze future performances, profits, claims, costs etc. We cover how to use ML models and clean data preparation strategies lagged features, lead features, trend, seasonality utilization, using moving average and statistical quantities as features to improve model performance and different models like linear regression, XGBoost, etc.
- Implement a machine learning pipeline that includes various encoding techniques (Label Encoding, One-Hot Encoding, and Ordinal Encoding), handles missing data through imputation and dropping values, and integrates model training using ColumnTransformer and pipeline methods.