FreakAnalytics

About

Hi! I’m Title (Nitiporn Jearrajinda), the creator of FreakAnalytics. I created this blog to share my passion for data analytics, business, and self-improvement, helping others make smarter, data-driven decisions in both work and life.

What I Share Here

  • Statistics & Data Analysis
  • Business & Strategy
  • Self-Improvement

Work Experiences

Evolt Technology
Leading EV charging solution provider in Thailand

  • Business Analyst – CEO Office (Nov 2022 – Aug 2024)
  • Project Management Officer – CEO Office (Sep 2024 – Present)

Projects

Loan Default Prediction with Cost-Based Threshold Tuning [GitHub]
Side Project

  • Developed a loan approval model using LendingClub dataset focused on profit maximization, not just accuracy.
  • Tuned decision threshold based on business cost function to identify the optimal cutoff with max profit of +146.6M THB.
  • Compared Logistic Regression vs XGBoost, selected Logistic for higher specificity (0.8150) to minimize financial risk from bad loan approvals.
  • Built reusable R scripts for data cleaning, modeling (caret), threshold tuning, and business cost visualization.
  • Applied LIME for model interpretability, enabling case-by-case explanation and audit-readiness.

Optimizing Bus and Charging Schedules Using the Harmony Search Algorithm
Evolt Technology

  • Developed a a metaheuristic-based scheduling algorithm using Harmony Search to minimize Total Cost of Ownership (TCO) for electric bus fleet operations
  • Designed customizable vehicle dispatch system supporting multi-station routing, rush hour adjustments, and cross-route fleet allocation
  • Integrated battery energy logic, including variable charging speed by battery level, and customizable charging station setup
  • Simulated real-world scenarios by incorporating CapEx and OpEx cost models, including vehicle, charger, installation, and on/off-peak electricity pricing
  • Enabled strategy-based scheduling, allowing users to switch between balanced operations and reinforcement-based planning for demand-centric deployments

Optimum Design for Steel Truss Structure Using Python
Senior Engineering Project

  • Developed a Python-based tool to optimize steel truss structures using Allowable Stress Design (ASD) from AISC standards.
  • Applied the Harmony Search algorithm to enhance truss member dimensions for structural integrity and cost efficiency.

Education

Chulalongkorn University 2018-2022
Bachelor of Engineering in Civil Engineering (GPA 3.12)

Ramkhamhaeng University 2022-2024
Dual Master’s Degree in Business and Public Administration (GPA 3.69)

Contact

LinkedIn : https://www.linkedin.com/in/nitiporn

Email : title.jea@gmail.com