Access the app here: https://hxiong6.shinyapps.io/dsproject/
Access the presentation here: https://drive.google.com/file/d/1L4v0GXLsv7oj_Gmnb7GMZ5sd_Ef5fBt1/view
flexdashboard; plotly; shiny;e1071; ggplot2; tidyverse; rvest;highcharter
This app displays the prediction of confirmed Covid-19 cases for the week ahead.
- The sidebar displays when the data in the app was last updated.
- Select state of which you are interested to see the trend of confirmed cases.
- A United States map shows the total confirmed cases of 50 states. Values are in k scale.
- In prediction part, the interactive plot will display the cumulative cases of Covid-19 of the state you choose since Jan,22,2020 and two predicted trends for the week ahead by SVM model and GLM model.
- In US data, the table shows the number of confirmed cases for last 7 days.
- The table summarize the confirmed coronavirus data around the globe.
- Overview of the project, disclaimer of the app, and references used through the project.
- Filter the data of the state user wants and combine them.
- Model learns from the date in which the number of confirmed cases is greater than 0.
- SVM: Use polynomial kernel and tune parameters of c and gamma by using trial and error. Implement the optimal SVR model to predict the confirmed case for the week ahead.
- GLM: Use quasipoisson model. Assume that the response variable Y (cumulative confirmed cases) follows a poisson distribution, and the logrithm of its expected value can be modeled by a linear combination of parameters.
- Data for prediction model https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_US.csv
- Data for USA https://www.worldometers.info/coronavirus/country/us/
- Data for globe. https://www.worldometers.info/coronavirus/
Ran Sui: SVM model; United states section;Flexdashboard making
Hao Xiong: GLM model;World section, resource and disclaimer section; Flexdashboard making