mkiang

## Imports ----
library(tidyverse)
library(genius)
library(here)
library(fs)
data("stop_words")
source("https://raw.githubusercontent.com/mkiang/airline_testing_strategies/master/code/mk_nytimes.R")

dir_create(here("data"))

mkiang

## Visualizing authorship networks: code for this blog post:
##  https://mathewkiang.com/2020/12/10/my-collaboration-network-for-2010-to-2020-other-plots

## Imports ----
library(scholar) # remotes::install_github("jkeirstead/scholar")
library(tidyverse)
library(fs)
library(here)

## Constants ----

mkiang

library(tidyverse)
library(ggrepel)
library(zoo)
source("https://raw.githubusercontent.com/mkiang/opioid_geographic/master/code/mk_nytimes.R")

## 2018 top causes of death:
## https://www.cdc.gov/nchs/fastats/deaths.htm
nchs_df <- tribble(~ cause, ~ deaths_2018,
        "1.Heart disease", 655381,
        "2. Cancer", 599274,

mkiang

## A quick R script for cleaning up your Twitter timeline. This assumes you
## have Twitter API creds but it probably isn't necessary. This is based on
## Chris Albon's Python script: 
##      https://gist.github.com/chrisalbon/b9bd4a6309c9f5f5eeab41377f27a670
##      
## In this example, I remove old tweets in two stages: (1) if tweets are 
## older than two years and have fewer than 100 likes and then (2) if tweets
## are older than 90 days and have fewer than 25. Retweets are removed after
## 90 days. Importantly, tweets (and retweets) that I liked are *always* 
## protected from deletion.

mkiang

library(tidyverse)
library(tidyquant)
source(
    "https://raw.githubusercontent.com/mkiang/disproportionate_prescribing/master/code/mk_nytimes.R"
)

getSymbols("^DJI")

DJI <- as_tibble(DJI) %>%
    mutate(

mkiang

## Visualizing authorship networks: code for this blog post:
## https://mathewkiang.com/2019/12/07/collaboration-network-from-2010-to-2019/

## Imports ----
library(scholar) # remotes::install_github("jkeirstead/scholar")
library(tidyverse)
library(fs)
library(here)

## Constants ----

mkiang

## Import packages
library(deSolve)

## Basic SIR using number of people (not proportion of pop)
SIR <- function(t, x, parms){
    with(as.list(c(parms, x)), {
        N = S + I + R
        dS <- -beta*S*I/N
        dI <- beta*S*I/N - r*I
        dR <- r*I

mkiang

install.packages("rgdal", dep = TRUE)
install.packages("reshape2", dep = TRUE)
install.packages("viridis", dep = TRUE)
install.packages("rgeos", dep = TRUE)
install.packages("Imap", dep = TRUE)
install.packages("network", dep = TRUE)
install.packages("DT", dep = TRUE)
install.packages("tidyverse", dep = TRUE)
install.packages("deSolve", dep = TRUE)
install.packages("manipulate", dep = TRUE)

mkiang

## Visualizing authorship networks: code for this blog post:
## https://mathewkiang.com/2018/06/17/my-collaboration-network/
library(scholar) # devtools::install_github("jkeirstead/scholar")
library(visNetwork)
library(tidyverse)

## Constants
MIN_TIME <- 60 * 5
MAX_TIME <- 60 * 30
PROJ_WEIGHT <- .8

mkiang

## For blog post here:  https://mathewkiang.com/2017/12/02/tldr-san-diego-weather-is-better-than-boston-weather/

## Imports ----
library(tidyverse)
library(GSODR)
library(ggmap)


## Define places we are interested in and google maps query ----
city_dict <- list(sandiego   = "san diego airport",