Yuzi-Liu's repositories
premium-user-conversion
The “freemium” business model — widely used by online services such as LinkedIn, Match.com, Dropbox, and music-listening sites — divides user populations into groups that use the service for free and groups that pay a fee for additional features. Given the higher profitability of premium subscribers, it is generally in the interest of company to motivate users to go from “free to fee”; that is, convert free accounts to premium subscribers. This project intends to analyze the data from an APP of an anonymized real music streaming company for potential insight to inform a “free-to-fee” strategy.
CyberSecurity_Malware_Analysis
The project's objecticve is to evaluate different classification models to predict malicious and benign websites, based on application layer and network characteristics.
Airline-Customer-Segmentation
Analyzed data of nearly 2 million Sun Country customers, used R and Tableau to decode and visualize data, segmenting customers to extract business intelligence.
Recommendations-for-Food
Recommend food for different users by predicting ratings for food recipes based on the rating history of each users
Movie-Review-Sentiment-Analysis
The primary objective for the project is to predict the sentiment of a movie review, using the dataset containing the text of the movie reviews from IMDB
Microsoft-Stock-Price-forecasting
The project objective is to forecast the stock price of Microsoft at the end of 2019 based on contextual factors.
Sales-Prediction
Delivering a Robust Sales Prediction Model in Under 2 Hours
content_based_recommender
Goal: Construct a recommender system. In the dataset, there are data of ingredients/ recipes, and interactions. Deliverable: Given a (user_id and recipe_id), predict the rating. Trying to minimize MSE (Mean Squared Error) The logic is: I want to recommend the dishes that have similar ingredients or recipes etc. with the dishes that a user rates high, which means the user likes the dish.
CTR-Analysis
This project broadly deals with location-based mobile marketing, using data from a location-based marketing agency which handles geo-fencing campaigns on behalf of advertisers. I choose to use a random sample for two campaigns of a single advertiser – AMC Theaters. The advertising impressions are inserted into the mobile app being used on the device. The data include the following elements: impression size (e.g., 320x50 pixels), app category (e.g., IAB1), app review volume and valence, device OS (e.g., iOS), geo-fence lat/long coordinates, mobile device lat/long coordinates, and click outcome (0 or 1).
social-network-analysis
The project involves social network analysis of product co-purchase data from Amazon. By finding information for product (books) itself and its neighbor (node neighbors, node degree, etc.), I was able to use Poisson regression to predict salesrank of all the books.