Arsyan Mardhi Arsanto
With a strong foundation in automation engineering, I derive insights from complex data and develop data-driven solutions to enhance decision-making and efficiency.
Contact Me
Full Name:
Arsyan Mardhi Arsanto
Email:
arsyansan@gmail.com
Phone:
+62 851-5623-5665
____________________________
Explore my resume and certifications here.
Data Scientist with a background in Automation Engineering, specializing in Machine Learning, Data Analysis, and Business Intelligence. Skilled in predictive modeling and interactive dashboards. Certified Associate Data Scientist (BNSP) with expertise in Python, SQL, and Looker Studio. Passionate about solving real-world problems through data-driven solutions.
Bachelor's of Applied Engineering - Diponegoro University | Semarang, December 2023
GPA: 3.32 / 4.00
Data Science Bootcamp - Digital Skola | July 2024
Graduated with Excellent Distinction.
.png)
.png)
• Analyzed feature importance to identify key churn factors like low engagement and late payments, guiding strategy.
• Developed a machine learning-based churn prediction model, utilizing SMOTE for balanced class distribution (1:1).
• Achieved 93.32% ROC-AUC and 85.50% accuracy using Gradient Boosting, the top model.
• Analyzed customer segmentation in car sales, identifying key factors like profession, spending score, and family size.
• Built and optimized classification models (Gradient Boosting, Random Forest, SVM) using GridSearchCV.
• Achieved 53.46% accuracy with Random Forest, the best-performing model.
• Enhanced loan eligibility prediction through feature engineering and optimization, increasing accuracy by 2.08%.
• Developed and optimized models (Random Forest, Logistic Regression) utilizing GridSearchCV for parameter tuning.
• Achieved 82.92% accuracy with Logistic Regression and deployed the top-performing model via Streamlit for improved accessibility.
• Performed EDA and preprocessing on a sarcasm detection dataset, refining headlines for analysis.
• Applied four NLP techniques (TF-IDF, Word2Vec, FastText) with four ML models, optimized via Bayesian Optimization.
• Naïve Bayes with TF-IDF (Removed Stopwords) achieved 84.73% of accuracy, highlighting strong feature importance with meaningful words.
• Analyzed trends from over 28 campaigns on Meta Ads and Google Ads, identifying key patterns to support strategic decision-making.
.
• Created 4 interactive dashboards to visualize campaign performance, ensuring clear and actionable insights for stakeholders.
• Monitored campaigns to evaluate performance metrics, identify cost-effective strategies, and provide actionable recommendations to support informed decision-making.
