Yuli Lin

I have a solid foundation in computer vision, data mining, and machine learning, especially in data mining and processing. Proficient in Python, SQL, and R, I have experience in data preprocessing, data mining, feature extraction, and backend development. I aim to leverage my skills in data analysis and data mining to contribute to impactful projects in a dynamic company.

Backend Developer, 07/2023 to 11/2023

Energy Storage Rights - Canberra, ACT

• Project Background: By capturing and integrating various types of energy resource data (such as wind energy, etc.), users can select a specific region or country on the website to obtain information about the types of energy and their quantities in that area.
• Data Crawling and Integration: Utilizing data scraping technologies from various open-source websites, efficient crawler programs are developed to gather diverse types of energy resource data globally. The information from different data sources is then integrated, ensuring the completeness and accuracy of the data.
• Data Processing and Analysis: The raw data collected is subjected to cleaning, format conversion, and standardization. Advanced data analysis methods are applied to uncover the distribution patterns, usage trends, and potential issues of energy resources.

Backend Developer, 02/2023 to 06/2023

Real-time monitoring of global food supply - Canberra, ACT

• Project Background: Initiated by two Australian government departments, this project primarily collaborates with the Australian Geoscience Bureau and the Food and Agriculture Organization. Its aim is to quickly assess the trends and patterns of crop production around the world.
• Data Scraping: Utilizing the Digital Earth Africa open-source website as the main source of data, including MODIS Terra LST, Global Precipitation Measurement, SMAP Daily, and other related datasets, to capture relevant NDVI data.
• Data Preprocessing: Clean the captured raw data, removing outliers, missing values, etc., to ensure data quality; Convert and standardize data from different sources to ensure consistency and comparability; Integrate the processed data to form a complete energy dataset.
• Data Analysis: Implemented Random Forest Regression to analyze past NDVI data and predict future monthly NDVI values. The steps included: Feature Engineering: Created time-based features such as month and year; Model Training: Split the data into training and testing sets, trained the Random Forest model using the training data; Prediction: Generated future NDVI predictions for the upcoming year by inputting the relevant features into the trained model, providing insights into expected crop production trends.
• Optimization: Implemented a strategy to dynamically adjust resolution based on the user's selected area size, reducing runtime by over threefold and keeping it within 5-10 minutes.