I'm Chonthichar Soythong (Pear)

Welcome to my Project Portfolio

🚀 Here you’ll find projects showcasing my achievements over the past 3 years, including internship work, personal projects, and academic assignments. My portfolio highlights data analysis skills from data cleaning and visualization to prediction and deployment for real-world scenarios.

Deep Learning and FastAPI Production in the Retail Industry (Norway and Sweden), achieved 98.8% Accuracy.

Integrated deep learning models into a web application using FastAPI to serve real-time predictions and data extraction for retail users

Visual Analytics - Data to Dashboard

Designed interactive dashboards in Tableau using sales/shipments across 6 European countries. Performed data cleaning, transformations, and built dashboards for sales performance, customer analysis, dispatch efficiency, using charts, maps, and advanced Tableau features.

Predictive Anlytics of Air Quality data, Hoursly Forecasting

Designed a forecasting model reducing prediction errors by 15% using VAR and ARIMA, with variable analysis, stationarity testing, and predictive analytics tools.

Surgical Instrument Detector

Built and deployed U-Net models on the Kvasir Instrument dataset, achieving over 90% accuracy in segmenting medical instruments. Conducted experiments on resolution, augmentation, data size, optimizers, and activation functions, with evaluation using Dice, IoU, precision, and recall.

Deep learning - CNN with CIFAR-10 and MNIST Dataset

Implemented a Multi-Layer Perceptron (MLP) for MNIST digit classification and a Convolutional Neural Network (CNN) for CIFAR-10 image classification using PyTorch. Achieved up to 99% accuracy on MNIST and 93% on CIFAR-10.

PetClinic Relational Database Design (MySQL)

Designed a normalized relational database for a veterinary clinic, with an ER diagram, schema, and SQL queries to manage pets, owners, vets, bookings, and appointments.

React-based map app

Built full-stack applications, including a React-based map app and a Python-Django platform.

The Instagram Scraper Application

Created a Python-based social media scraper with Selenium and KivyMD, automating data collection for public accounts.

📧 chonthichar1995@hotmail.com

📞 +(47)-92213789

📍 Based in Oslo, Norway