Project 1: Classification Model for Production Floor (Cycle Time)

• 1. Extracted outliers in the original data from the database by applying the Interquartile range (IQR) and confidence interval method to increase the data accuracy from 70.6% to 91.1%.
• 2. Analyzed and maximized manpower using the optimal strategy of arranging operators to assemble different models, resulting in a 17% increase in completion efficiency per order on average.
• 3. Envisioned the data of each operation by using Python.
• 4. Converted the monitoring mechanism system, which improved efficiency by 20%.

• 1. Created a video to visualize the analyzed data by applying the Geopandas model and detected the trends of wildfires in the year 2022 to create readable data for the research team.
• 2. Predicted the daily acres of wildfire impact by applying the Sklearn model.
• 3. Collected the 5+ factors that may cause the wildfires, and created an algorithm to calculate the risk level of a specific area by factors.

• 1. Developing an image of an unmeasurable object (Digital Twin) enabling the object to be measured and monitored on the dashboard.
• 2. Applied the OpenCV package to do image detection, to identify the object's main part, thus avoiding the problem of an uncentered object, or the camera angle being tilted.
• 3. Designed an algorithm by extracting the feature images from cameras and coordinated data from sensors to build the digital mapping in order to display the physical conditions of the object.