math370-final

370code.Rmd

---
title: "370 Work"
author: "Ryan Grossman"
date: "4/20/2021"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r}
library(tidyverse)
library(covidcast)
setwd("~/Documents/public health 460/data")
us_locations <- read_csv("us_locations.csv")

covid_cases <- covidcast_signal(data_source = "jhu-csse",
signal = "confirmed_incidence_num",
start_day = "2021-01-01", end_day = "2021-04-19",
geo_values = )

vaccine_acceptance <- suppressMessages(
covidcast_signal(data_source = "fb-survey", signal = "smoothed_covid_vaccinated_or_accept",
                 start_day = "2021-03-22", end_day = "2021-04-19",
                 geo_type = "county")
)

mask_wearing <- suppressMessages(
covidcast_signal(data_source = "fb-survey", signal = "smoothed_wearing_mask_7d",
                 start_day = "2021-01-01", end_day = "2021-04-19",
                 geo_type = "county")
)

travel <- suppressMessages(
covidcast_signal(data_source = "safegraph", signal = "full_time_work_prop_7dav",
                 start_day = "2021-01-01", end_day = "2021-04-19",
                 geo_type = "county")
)

symptom_search <- suppressMessages(
covidcast_signal(data_source = "google-symptoms", signal = "sum_anosmia_ageusia_smoothed_search",
                 start_day = "2021-01-01", end_day = "2021-04-19",
                 geo_type = "county")
)

doctor_visits <- suppressMessages(
covidcast_signal(data_source = "doctor-visits", signal = "smoothed_adj_cli",
                 start_day = "2021-01-01", end_day = "2021-04-19",
                 geo_type = "county")
)

covid_like_symptoms <- suppressMessages(
covidcast_signal(data_source = "fb-survey", signal = "smoothed_cli",
                 start_day = "2021-01-01", end_day = "2021-04-19",
                 geo_type = "county")
)


```


```{r}
num_covid_cases <- covid_cases %>% 
group_by(geo_value) %>% 
summarize(num_covid_cases = sum(value))
num_covid_cases

avg_vaccine <- vaccine_acceptance %>% 
group_by(geo_value) %>% 
summarize(avg_vaccine = mean(value))
avg_vaccine

avg_mask <- mask_wearing %>% 
group_by(geo_value) %>% 
summarize(avg_mask = mean(value))
avg_mask

avg_travel <- travel %>% 
group_by(geo_value) %>% 
summarize(avg_travel = mean(value))
avg_travel

avg_symptomsearch <- symptom_search %>% 
group_by(geo_value) %>% 
summarize(avg_symptomsearch = mean(value))
avg_symptomsearch

avg_docvisits <- doctor_visits %>% 
group_by(geo_value) %>% 
summarize(avg_docvisits = mean(value))
avg_docvisits

avg_covidlikesymptoms <- covid_like_symptoms %>% 
group_by(geo_value) %>% 
summarize(avg_covidlikesymptoms = mean(value))
avg_covidlikesymptoms
```


```{r}
data <- left_join(x = num_covid_cases, y = avg_vaccine)
data <- left_join(x = data, y = avg_mask)
data <- left_join(x = data, y = us_locations)
data <- left_join(x = data, y = avg_travel)
data <- left_join(x = data, y = avg_symptomsearch)
data <- left_join(x = data, y = avg_docvisits)
data <- left_join(x = data, y = avg_covidlikesymptoms)
data <- mutate(data, fraction_infected = (num_covid_cases)/(population))
data <- filter(data, !is.na(avg_vaccine))
data <- filter(data, !is.na(avg_mask))
data <- filter(data, !is.na(avg_travel))
data <- filter(data, !is.na(avg_symptomsearch))
data <- filter(data, !is.na(avg_docvisits))
data <- filter(data, !is.na(avg_covidlikesymptoms))
```


```{r}
mod1 <- lm(fraction_infected ~avg_vaccine, data = data)
mod2 <- lm(fraction_infected ~avg_mask, data = data)
summary(mod1)
summary(mod2)
```


```{r}
ggplot(data, aes(avg_vaccine, fraction_infected)) + geom_point() + geom_smooth(method = 'lm')
```


```{r}
e.mod1 <- mod1$residuals
e.mod2 <- mod2$residuals
#ggplot(data, aes(avg_vaccine, e.mod1)) + geom_point(alpha = 0.4)
ggplot(data, aes(avg_mask, e.mod2)) + geom_point(alpha = 0.4)

model <- lm(fraction_infected ~ avg_vaccine + avg_travel + avg_symptomsearch + avg_mask + avg_docvisits + avg_covidlikesymptoms, data = data)

#mse(model, data)

library(olsrr)

summary(model)

ols_step_forward_p(model)


```

Credit and author: Ryan Grossman