Highlights from Github Portfolio
Recommender Systems for MovieLens
- Several popular & bleeding edge approaches
- top-N Rankings VERSUS Ratings prediction
- Challenges of scaling, cold-start, stoplist, filter bubbles, trust etc
Anti Money Laundering (AML) / Fraud Detection
- Creating candidate variables from domain knowledge for adversarial learning
- Treating class imbalance, missing values imputation, model selection using cross-validation, generalization performance for recall
- Explainable models with Shapley
Sarcasm Detection in News Headlines
- Robust Text Pre-processing Steps
- Feature Engineering : Bag of Words & Semantics based using word2vec
- Feature Importance
RShiny App for Video Games / Movies / TV Shows
- RShiny interactive app
- For sales and ratings
- Sales by year, platform and genre
Object Oriented Programming in Python
- Classes, Objects, Methods
- OOPs Concepts - Encapsulation, Abstraction, Inheritance, Aggregation, Composition
- Other Key Building Blocks
⚽ Football Players’ Popularity Analysis through Twitter Streaming API
- Twitter Live Streaming with Spark
- Real Time Sports Analytics
- Cumulative Count & Sentiment
Time Series Forecasting with Exogenous Features
- Forecasting with Exogenous Features
- Prophet vs SARIMAX
- Rolling Window Cross-Validation
Apache Spark with Scala
- Scala Functional Programming
- Variables, Flow Control, Functions, Data Structures
- ML in Spark using Scala
DevOps & Cloud Essentials
- DevOps Methodology & Lifecycle
- Agile, DevOps, CI/CD
- Key DevOps Tools & Cloud Services
Hugging Face 🤗
- Hugging Face Universal Interface
- Popular NLP Tasks Notebooks (using Transformers)
- Use Case Examples & Evaluation Metrics
⚽ Football History Time Lapse
- Analyzing International Fottball Matches
- Geo Plot GIF using R
- Football History Time Lapse (years 1872 - 2018)