Yixi K

- Research Program Mentor

Bio

Project ideas

Interactive COVID data dashboard

This is a sample project I mentored a high school student through at the beginning of COVID. We took state published raw data and build an interactive dashboard very similar to the Washington state's dashboard (which launched several months after ours) through Power BI. https://doh.wa.gov/emergencies/covid-19/data-dashboard

Regression model to predict house sale price

Using regression model leveraging house characteristics (bedrooms, garage, fireplace, pool, porch, etc.), location (neighborhood), lot information (zoning, shape, size, etc.), ratings of condition and quality to predict the house sale price.

Interactive calculator to calculate buy vs rent for housing

Build an interactive calculator to calculate buy vs rent for housing based on different inputs. Another idea is to build a calculator to analyze savings rate and retirement date.

Social media change the marketing industry for organisations with smaller marketing budgets

Research paper: How did social media change the marketing industry for organisations with smaller marketing budgets (non- profits, start-ups, small business) and help reach more specific demographics in order to increase customer relationships and sales? As well as how has the recent changes in data privacy impact above.

Regression Model for Solar Radiation/UV Estimation

* Dataset: Datasets containing solar irradiance measurements and weather parameters like cloud cover, sunshine duration, etc. * Task: Build a model to estimate solar radiation levels based on these weather variables. This can be useful for solar energy applications/UV.

Regression Model for Water Quality Prediction

* Dataset: Find data on water quality parameters (e.g., pH, dissolved oxygen, turbidity) for various water bodies (rivers, lakes, etc.). Government environmental agencies often collect and share this type of data. * Task: Use regression to predict water quality parameters based on factors like upstream land use, rainfall, temperature, or other environmental variables.

Regression Model for Air Pollution Prediction

* Dataset: Look for air quality data (e.g., pollutant concentrations) along with meteorological data (e.g., temperature, wind speed). Many cities and regions have air quality monitoring networks. * Task: Build a model to predict air pollution levels based on weather conditions and other relevant factors. This could be useful for public health alerts or urban planning.

Regression Model for COVID-19 Case Prediction

* Dataset: The Johns Hopkins University COVID-19 dataset (available on GitHub or Kaggle) is a great starting point. It contains daily confirmed cases, deaths, and recoveries for various countries/regions.   * Task: Use regression to predict daily new cases based on factors like previous case numbers, testing rates, population density, etc. This can be done for a specific country/region or globally.

Regression Model for Vaccination Rate and Case Reduction

* Dataset: Combine the Johns Hopkins dataset with vaccination data from Our World in Data or similar sources. * Task: Analyze the relationship between vaccination rates and changes in COVID-19 case numbers. Regression can help quantify this relationship and potentially predict case reductions based on vaccination progress.

Regression Model for Hospitalization Prediction

* Dataset: Local or regional health department data often includes hospitalization numbers. * Task: Predict hospitalizations based on factors like new cases, vaccination rates, age demographics, etc. This could help healthcare systems anticipate resource needs

Regression Model for College Admission Prediction

* Dataset: Find datasets containing information about college applicants, such as GPA, standardized test scores, extracurricular activities, demographic information, and admission decisions. You can find these on Kaggle or through educational institutions that share anonymized data. * Task: Use logistic regression (a type of classification) to predict the probability of college admission based on applicant characteristics. This can help students understand their chances of getting into different schools.

Regression Model for High School Dropout Prediction

* Dataset: Look for datasets containing student information such as attendance, grades, demographics, and whether they dropped out of high school. These can be obtained from educational institutions or research organizations. * Task: Use regression to identify factors that contribute to high school dropout rates. This can help schools implement interventions to support students at risk of dropping out.

Regression Model for Student Performance Prediction

* Dataset: Gather data on student grades, attendance, study habits, and demographic information. Schools often have this data internally. * Task: Build a regression model to predict student performance on standardized tests or in specific courses. This can help teachers identify students who may need additional support.

Regression Model for Impact of School Resources on Student Outcomes

* Dataset: Combine data on student outcomes (e.g., test scores, graduation rates) with information on school resources (e.g., teacher-student ratios, funding per student). You can find these datasets from government education agencies or research institutions. * Task: Use regression to analyze the relationship between school resources and student outcomes. This can inform education policy and resource allocation decisions.

Regression Model for Financial Aid Eligibility Prediction

* Dataset: Collect data on student financial information (e.g., family income, assets) and financial aid awards. Educational institutions or government agencies may have this data available. * Task: Develop a regression model to predict financial aid eligibility based on student financial circumstances. This can help students understand their potential financial aid options.

Coding skills

Languages I know

Teaching experience

Credentials