Jun Du

 Operate the Analysis and Prediction of Customer Churn 

• Explored 9,000+ customer records and conducted Explorative Data Analysis of consumers with Python.
• Identified high lifetime value customers and visualized their distinctive characteristics.
• Output data analysis reports can obtain the characteristics of users with high turnover rates, which can reduce the company's user turnover rate by 7% and increase the company's profit by 35%.

Kaggle's Negative Comment Data Classification 

• Managed 3,000 Kaggle malicious comment data records, addressing data imbalance through undersampling and oversampling techniques.
• Implemented LSTM and BERT models to achieve classification accuracy rates of 96% and 100%.
• Used Streamlit to build a data scientific APP based on Bert’s model, as it can classify comments on the web page and improve the efficiency of follow-up work management by 40%.

P2P Credit platform business data analysis

• Cleaned and visualized 18,000 data and conducted data preprocessing, feature selection and scaling with Python.
• Analyzed the capital deposit, late payments, and borrowing rates, the proportion of the user's repayment habits, and visualized specific characteristics with Python (Matplotlib).