Insight-Sogang-Univ/insight-13th
May 2025 – Jun 2025
2 Months active
Languages Used
Jupyter NotebookPythonSQL
Technical Skills
ADF TestAssociation Rule MiningBERTCatBoostClassificationCollaborative Filtering
zzuhi
PROFILE
Zzuhi
Juhui worked on the Insight-Sogang-Univ/insight-13th repository, delivering eight end-to-end AI and machine learning features over two months. Their work included building an employee leave prediction model, collaborative filtering recommendation system, and deep learning pipelines for digit classification and time-series forecasting. Juhui applied techniques such as ARIMA modeling, LSTM-based forecasting, and association rule mining, using Python, PyTorch, and SQL for data loading, preprocessing, model training, and evaluation. The solutions addressed business needs in HR analytics, e-commerce, and forecasting, with reproducible, well-documented workflows that demonstrated depth in both statistical and deep learning approaches without reported production bugs.
Overall Statistics
Feature vs Bugs
100%Features
Repository Contributions
9Total
Bugs
0
Commits
9
Features
8
Lines of code
27,322
Activity Months2
Your Network
1029 people
Work History
June 2025
2 Commits • 2 Features
Jun 1, 2025June 2025: Delivered two time-series forecasting capabilities in Insight-Sogang-Univ/insight-13th, enabling both deep-learning and statistical approaches for business forecasting. Implemented an end-to-end LSTM-based forecasting pipeline with sliding-window inputs (data loading, differencing/preprocessing, a custom PyTorch Dataset, and an LSTM+fully connected model), with evaluation using Pearson and Spearman correlations. Added ARIMA-based Time Series Analysis Notebook for Seoul's monthly temperatures, including data loading, preprocessing, visualization, decomposition, stationarity testing (ADF/KPSS), and interpretation of ACF/PACF plots for ARIMA parameter selection. No major bugs reported in this period; efforts focused on feature delivery, reproducibility, and tooling. Technologies demonstrated include PyTorch, ARIMA, Jupyter notebooks, and standard time-series diagnostics.
2 Commits • 2 Features
Jun 1, 2025June 2025: Delivered two time-series forecasting capabilities in Insight-Sogang-Univ/insight-13th, enabling both deep-learning and statistical approaches for business forecasting. Implemented an end-to-end LSTM-based forecasting pipeline with sliding-window inputs (data loading, differencing/preprocessing, a custom PyTorch Dataset, and an LSTM+fully connected model), with evaluation using Pearson and Spearman correlations. Added ARIMA-based Time Series Analysis Notebook for Seoul's monthly temperatures, including data loading, preprocessing, visualization, decomposition, stationarity testing (ADF/KPSS), and interpretation of ACF/PACF plots for ARIMA parameter selection. No major bugs reported in this period; efforts focused on feature delivery, reproducibility, and tooling. Technologies demonstrated include PyTorch, ARIMA, Jupyter notebooks, and standard time-series diagnostics.
June 2025
May 2025
7 Commits • 6 Features
May 1, 2025In May 2025, delivered a multi-domain AI/ML slate for Insight-Sogang-Univ/insight-13th, spanning HR analytics, recommendations, e-commerce insights, deep learning experiments, NLP, and time series. End-to-end work from data loading to model evaluation and iteration enabled tangible business value: proactive workforce planning, personalized recommendations, and analytics readiness for production experiments. Key features include: Employee Leave Prediction Model, Collaborative Filtering Recommendation System, MNIST Digit Classification (MLP) Iteration 2, NLP with BERT/GPT and RAG Setup, and Time Series Analysis with STL and stationarity tests.
May 2025
7 Commits • 6 Features
May 1, 2025In May 2025, delivered a multi-domain AI/ML slate for Insight-Sogang-Univ/insight-13th, spanning HR analytics, recommendations, e-commerce insights, deep learning experiments, NLP, and time series. End-to-end work from data loading to model evaluation and iteration enabled tangible business value: proactive workforce planning, personalized recommendations, and analytics readiness for production experiments. Key features include: Employee Leave Prediction Model, Collaborative Filtering Recommendation System, MNIST Digit Classification (MLP) Iteration 2, NLP with BERT/GPT and RAG Setup, and Time Series Analysis with STL and stationarity tests.
Activity
Loading activity data...
Quality Metrics
Correctness76.6%
Maintainability75.6%
Architecture75.6%
Performance67.8%
AI Usage28.8%
Skills & Technologies
Programming Languages
Jupyter NotebookPythonSQL
Technical Skills
ACF/PACF AnalysisADF TestARIMA ModelingAssociation Rule MiningBERTCatBoostClassificationCollaborative FilteringCosine SimilarityData AnalysisData DecompositionData PreprocessingData VisualizationDeep LearningDifferencing
Repositories Contributed To
Overview of all repositories you've contributed to across your timeline