Description

Why are some people seen as effective leaders and others are not? Are there any behaviors or characteristics that can help us quantify what an effective leader looks like?

I used the data from a large survey of employees and their direct manager. This data contains individual items and the scale score for those items. The main goal is to use individual items, scale subscores, and/or scale scores to explain the effect variable.

I used R to apply hypothesis testing, power analysis, bootstrap resampling and robust standard error.

Description

Music contains so much information. We can not only compare all the hit songs to conclude the popular music trend, but also analyse the song behaviour of a particular person and get to know more about him/her through just a small music application.

Spotify is one of the top music streaming service around . With 35+ million songs and 170+ million monthly active users, musicians can spread their music and reach the audience easily. Also, it is an ideal platform for users to find their favorite music quickly and conveniently due to its large catalog, collaborative playlists, podcasts, and other attractive features. On the app, music can be browsed or searched via various parameters — such as artists, album, genre, playlist, or record label. Users can create, edit, and share playlists, share tracks on social media, and make playlists with other users.

This project is divided into two parts. The goal in Part I is to analyze what song characteristics would affect its popularity. To achieve this, I built up a linear mixed model using R. The goal in Part II is to create a like song prediction system for a specific spotify user. I used R to apply logistic regression model and deployed it in a R Shiny App

Description

Like many traditional media companies, Daily Media Blitz, a media company based in Utah, has been in decline for over a decade. Much of their revenue is derived from selling advertising— though that has changed somewhat in the last few years since a digital agency was created to sell digital advertising and services (SEO, website design, social media management, etc.) The company hoped to replace the lost print revenue with new streams of revenue coming from the digital agency.

It is a team project completed by me, Weiyi Zhao, Dijia Tang, Zhijing Zhao and Rebecca Liu. We used Tableau to provide daily in-time dashboards using the daily transactional information (1/1/16 – 5/1/18). We have conducted analyses on health of the sales operation, customer success, salesperson, publication portfolio, and preprint rate.

Description

Due to the ongoing economic and humanitarian crisis in Venezuela, approximately 1.4 million Venezuelans have fled to Colombia in recent years. In this visualization report, we are going to provide analysis on where majority of Venezuelan refugees and/or asylum seekers are currently distributed around the globe and how the inflow of Venezuelans has impacted Colombia's economy and societal development.

It is a team project completed by me, Weiyi Zhao, Zhijing Zhao and Yunqi He for the CGDV Data Viz Hackathon 2020. We used Tableau to communicate a story of general situation, economy under crisis, the voices in Venezuela migration crisis and the future of migration, and provide several recommendations for venezuelans both residing and migrating to Colombia.

Description

Learning the sentiments and topics associate with reviews of employees provided to a third party reviews website like Glassdoor might be useful intelligence for HRs and management board at the company in understanding its perceived strengths and weaknesses from an unfiltered employee perspective.

The analysis focuses on the Glassdoor employee reviews of 4 organizations. The reviews are broken into 3 separate parts: pros, cons, and advice. Using these reviews, sentiment analysis and topic modeling are conducted to find out the specific topics within a corpus of documents and predict the rating. In addition, a CNN(convolutional neural network) model was built up to predict the ratings based on the given reviews. All the works are performed in R.

Description

Earnings call is a conference call among the management of a public company, analysts, investors, and the media to discuss the company’s financial situation during a given reporting period, such as a quarter or a fiscal year.

In this analysis, I dig into the WWE earnings call transcript and perform sentiment analysis with various lexicons. In addition, I scraped WWE’s stock prices using the Alpha Vantage API to analyze the relationship between closing prices and the sentiment scores and try to find out the investment behavior patterns based upon a call. I also created an interactive visualization showing time series trend line under different sentiment scores with plotly.

All the works are performed in R.

Description

The analysis of credit risk and the decision making for granting loans are important tasks for financial institutions. With historical data, we are able to train a model and thus accurately predict future outcomes.

After the data cleansing and feature selection (Recursive Feature Elimination & Pearson Correlation) phase, I built up a logistic regression model to predict the loan default likelihood. Also, I used grid-search cv to improve the model performance and used cross-validation to further validate the model. I also consider the data leakage issue and consider whether all the default data are created equal. All the works are performed in Python and the dataset comes from Kaggle.

Description

Flight delays and cancellations are typical problems all of us face when traveling. The airports and airline industry try to minimize the impact on the customers and improve their experience. On the other the airline industry also optimizes the flights to fly through their network of cities.

In this project I worked with dataset compiled by Kaggle providing summary information on the number of on-time, delayed, canceled, and diverted flights published by DOT’s montly Air Travel Consumer Report for the year 2015.

With the matplotlib and plotly package in Python, I created various visualizations in order to get some insights and recommendationss on the flight delays and cancellation situation.