rolling_median_notebook.Rmd

---
title: "R Notebook"
output: html_notebook
---

```{r}
library(dplyr)
library(ggplot2)
library(tidyr)
library(lubridate)
```


```{r}
x <- read.csv('athenaResults.csv')
y <- x %>% group_by(as_of_when) %>% summarise(total_today = sum(today_count))
n <- 5

median_roll <- numeric(length(y$as_of_when) - n)
mean_roll <- numeric(length(y$as_of_when) - n)
for(i in seq_along(median_roll)){
  median_roll[i] <- median(y$total_today[i:(i+n)])
  mean_roll[i] <- mean(y$total_today[i:(i+n)])
}
median_roll = c(numeric(n), median_roll)
mean_roll = c(numeric(n), mean_roll)
y$median <- median_roll
y$mean <- mean_roll
y$as_of_when <- as.Date(y$as_of_when)
```

```{r}
LAST_N_DAYS <- 30
bisect_date <- as.Date(today(), format="%Y-%m-%d") - LAST_N_DAYS

#select(y, -total_today) %>%
y %>%
  gather(variable, value, -as_of_when) %>%
  filter(as_of_when >= bisect_date) %>%
  ggplot(aes(x=as_of_when, y=value, color=variable)) +
  geom_line() #+ scale_y_log10()

```