Data Analyst · Data Scientist · Data Engineer, building end-to-end data projects, from raw data to deployed models to reproducible pipelines. Currently expanding my portfolio across all three tracks after completing a Data Science Trainee program with Pusdatin, Ministry of Primary and Secondary Education (Kemendikdasmen) Indonesia.
I care about building things that hold up under scrutiny, a lot of my recent work has been about auditing my own past results and fixing subtle issues (data leakage, embedding mismatches, evaluation shortcuts) that inflate reported numbers without anyone noticing.
- 📊 Data Analysis — exploratory analysis, statistics, turning raw data into decisions
- 🤖 Data Science / Machine Learning — from classical ML to deep learning (PyTorch, TensorFlow), with a specialty in NLP for Bahasa Indonesia
- 🔧 Data Engineering — clean, reproducible data pipelines and workflows
- 🔍 Research integrity & reproducibility — revisiting completed work to find and fix methodological gaps, not just chasing the highest accuracy number
I'm actively building a portfolio across Data Analyst, Data Science, and Data Engineer project tracks — this repo (and the list below) will keep growing as each one ships.
Klasifikasi Hoax Berita Politik Menggunakan Algoritma Long Short-Term Memory (LSTM) dengan Penambahan Fitur Embedding Global Vector (GloVe)
Sunan, R.A., Erliawan K., H.F., & Aditya, C.S.K. (2024) — JEPIN (Jurnal Edukasi dan Penelitian Informatika), Vol. 10 No. 2, SINTA 2
| Track | Project | Description |
|---|---|---|
| Data Science / NLP | 🔍 klasifikasi-hoax-lstm-glove | Archived, as-published code from my JEPIN paper above, kept transparent, including a documented post-publication finding about the embedding setup. |
| Data Science / NLP | 🛠️ klasifikasi-hoax-lstm-revisited | Independent revision of the project above: rebuilt in PyTorch with FastText Indonesia (subword-aware embeddings), fixed a data-leakage bug in the original imbalance-handling step, with a full before/after results comparison. |
| Data Analysis | ⚖️ Dataset_Narkotika | Curated dataset of narcotics/psychotropics court decisions from Pengadilan Negeri Rantau Prapat, sourced from Indonesia's Supreme Court decision directory, for legal research and search-engine development. |
| Data Analysis / Data Science | 📊 ULT Complaints Analytics (private — pending data-use authorization) | Topic modeling (LDA) + sentiment analysis (IndoBERT) on a government complaints platform, built during my Data Science Trainee program. Presented internally; not yet public pending institutional approval. |
| Data Engineer | 🚧 In progress | A data pipeline/engineering project is next on my roadmap, check back soon. |
(This table grows as each project ships — see the roadmap note above.)
- 📧 Email: rohidajisunan185@gmail.com
- 💼 LinkedIn: Rohid Aji Sunan