Alexander Tiopan

AI_Email_Crafter. Effortlessly generate personalized, VC-ready email drafts using AI and save them directly into Gmail — a smart outreach assistant for founders, sales, and BD teams.

Project-Progress-Monitoring. Interactive Google Sheets-based project tracker with Gantt chart, daily reporting, activity selector, and issue photo upload to Drive.

Brent_Oil_Price_Prediction. Using RNN with LSTM neural networks, this project forecasts Brent oil prices by preprocessing historical data, training the model with dropout, and evaluating for accurate predictions.

PV-Power-Plant-Calculation. Interactive Google Sheets template for PV power plant feasibility calculation — includes OpenStreetMap location picker, NASA GHI data integration, and battery sizing simulation.

DataPulse. A real-time CDC (Change Data Capture) simulation dashboard using SQLite to mimic ETL pipelines and visualize live sales data. Built with Streamlit and Plotly, includes multi-user login support.

Compressor_Control_Offline_RL. The project models a compressor machine to predict parameter changes, ensuring stable pressure. It preprocesses data, employs ensemble methods, and utilizes reinforcement learning.

Streamlit_App_GIS_Clustering_FCM_PSO. A Streamlit-based web application for geospatial clustering using Fuzzy C-Means (FCM) and Particle Swarm Optimization (PSO), visualized on an interactive map with support for shapefiles in .zip format. Includes a blog module for adding and managing user-generated articles.

Streamlit_App_Cheat_Detection. CheatDetection is a Streamlit-based web application designed to detect cheating behavior during online exams. Utilizing computer vision and deep learning techniques (e.g., face detection, gaze tracking), the app monitors test-takers in real-time, generates violation reports, and allows proctors to analyze visual evidence through a user-friendly int

PDF_Malware_Detection. The project enhances GARUDA's cybersecurity by employing deep learning to classify malware in scholarly resources, utilizing Python libraries and evaluation metrics.

Kemenkeu_Bandwidth_Prediction. Forecast time series using ARIMA, employing exploratory data analysis, data preprocessing, stationarity checks, AR and MA modeling, differencing, and Python libraries.

ArbiDex. ArbiDex is a smart prediction market arbitrage dashboard that detects mismatches between similar Manifold markets using fuzzy logic, complete with actionable insights, weekly trend tracking, and historical database.

SmartScan-AI. SmartScan-AI is a Streamlit app for invoice & PO extraction, matching, and AI-powered document Q&A.

GeoTrak. Interactive Streamlit-based GIS explorer to visualize and drill into UK catchment shapefile layers with map clicks and area calculation.

capitalized. Capitalized is an all-in-one financial modeling platform to value equities, bonds, options, and simulate project financing — powered by real-time market data and interactive insights.

Indonesian_Spelling_Corrector. Developing Indonesian spelling correction using RNN model on root word dataset. Preprocess with tokenization and root form conversion, train, evaluate, and potentially improve deployment.

Streamlit_App_Clustering_Training_Customer. Interactive customer segmentation app using KMeans and KMedoids clustering on alumni training data from POLTEKPEL Banten, powered by Streamlit.

Streamlit_App_Hospital_Revenue_Predictor. An interactive Streamlit app for clustering insurance companies and forecasting hospital revenue using linear regression and time series modeling.

Classify_Fish_Species. This project aims to classify fish species using morphological features extracted from fish images, including dissimilarity, correlation, and energy.

Streamlit_App_Kemenkeu_Bandwidth_Predictor. A Streamlit-based forecasting app to analyze and predict daily bandwidth utilization for Indonesia’s Ministry of Finance using ARIMA and Holt-Winters methods.

Fracture_Leg_Detection. Developing a fracture detection system for leg radiographs using VGG16 pre-trained CNN, aiming to enhance diagnosis efficiency in clinical settings.

Clustering_Mobilphone_Sales. Cluster mobile phones by attributes for market segmentation. Employ K-means, preprocessing, and visualization tools in Python for analysis and insights.

Clustering_Training_Customer. Analyze Banten Marine Polytechnic cadet/alumnus profiles with age, recency, frequency, and monetary methods, clustering via K-means algorithm from scratch.

Hospital_Revenue_Prediction. Clustering hospital revenue from insurance claim payment using timeseries clustering method.

GE_Visualization. Utilize the GE McKinsey Matrix methodology to assess product strategic positioning within the market using Python tools.

American_Option_Price. Predicts the fair value of Company XYZ's during participation rights in the X project using American Binomial Options Method.

Gold_Price_Prediction. A custom Radial Basis Function (RBF) neural network is implemented to predict gold prices using time series data.

Stock_Trend_Prediction. Employing TOBA stock data, the random forest algorithm predicts price trends, reaching a 70.2% accuracy post hyperparameter tuning and feature selection.

Clustering_Digital_Wallet_Users. This project aims to cluster digital wallet users based on their survey responses using the KPrototypes algorithm, which handles both numerical and categorical data, resulting in three distinct clusters characterized by different demographic and behavioral attributes.

District_Graph_Coloring. Apply Greedy and Welch Powell algorithms to assign colors to map districts ensuring adjacent districts have different colors.

Predicting_Overweight. The project utilizes Support Vector Regression to predict individual weights from obesity data, contributing to personalized healthcare strategies.

Classify_Wine_Class. The project implements Adaboost from scratch to classify wine classes using features from the wine dataset.

Kompas_News_Hoax_Detection. Automates fake news identification in Kompas articles, comparing Random Forest and Convolutional Neural Network models for text classification.

Pytorch_to_ONNX_Converter. Install packages, define CNN in PyTorch, load weights, export to ONNX, fix input name error, save model.