TwitterWeatherVane/WeatherVane.R at main · ColinJoMurphy/TwitterWeatherVane · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
# Load required packages
library(rtweet)
library(tidytext)
library(httpuv)
library(dplyr)
library(tibble)
library(stringi)

# Set weather word mappings
keywords <- c('sunny',
           'sunshine',
           'partly',
           'cloudy',
           'overcast',
           'rainy',
           'rain',
           'raining',
           'showers',
           'snow',
           'snowing',
           'snowy',
           'windy',
           'foggy',
           'thunder and lighting')

types <- c('Sunny',
           'Sunny',
           'Partly Cloudy',
           'Cloudy',
           'Overcast',
           'Rainy',
           'Rainy',
           'Rainy',
           'Showers',
           'Snowy',
           'Snowy',
           'Snowy',
           'Windy',
           'Foggy',
           'Thunder and Lighting'
           )

weatherkey <- tibble(keywords, types)

# Set up access to bot account
auth <- rtweet_bot(api_key = Sys.getenv('WEATHER_VANE_BOT_API_KEY'),
                   api_secret = Sys.getenv('WEATHER_VANE_BOT_API_SECRET'),
                   access_token = Sys.getenv('WEATHER_VANE_BOT_ACCESS_KEY'),
                   access_secret = Sys.getenv('WEATHER_VANE_BOT_ACCESS_SECRET'))
auth_as(auth)

# Pull tweets related to weather and are from Sacramento, CA area
t <- search_tweets(q = 'weather', n = 50, type = 'recent', include_rts = F, geocode = '38.575764,-121.478851,25mi')

# Get today's date and format
today <- Sys.Date()
textdate <- format(today, '%A, %B%e') %>% tolower() %>% stri_replace( '', fixed = ',')

# Pull tweets and analyze to find the weather
weather <- t %>%
  tibble() %>%
  unnest_tweets(word, full_text, drop = FALSE) %>%
  unnest_ngrams(grams, full_text, drop = FALSE) %>%
  group_by(id_str) %>%
  filter(any(word == 'today') | any(grams == textdate)) %>%
  filter(strftime(created_at, format = '%Y-%m-%d') == Sys.Date()) %>%
  ungroup() %>%
  mutate(word = ifelse('partly' %in% word & word == 'cloudy',
                       'partly',
                       word)) %>%
  inner_join(weatherkey, by = c('word' = 'keywords'), keep = T)

# Find and format high temps
highs <- weather %>%
  filter(grepl('^high of', grams)) %>%
  select(grams) %>%
  unnest_tokens(char, grams) %>%
  filter(grepl('^[0-9]', char)) %>%
  deframe() %>%
  stri_replace_all_regex('f', '') %>%
  unique() %>%
  as.numeric() %>%
  sort() %>%
  as.character() %>%
  stri_c('f') %>%
  stri_c(collapse = ' or ')

# Find and format low temps
lows <- weather %>%
  filter(grepl('^low of', grams)) %>%
  select(grams) %>%
  unnest_tokens(char, grams) %>%
  filter(grepl('^[0-9]', char)) %>%
  deframe() %>%
  stri_replace_all_regex('f', '') %>%
  unique() %>%
  as.numeric() %>%
  sort() %>%
  as.character() %>%
  stri_c('f') %>%
  stri_c(collapse = ' or ')

# Generate message
message <- weather %>%
  select(types) %>%
  unique() %>%
  deframe() %>%
  stri_c( collapse = ' or ') %>%
  paste0('Twitter says Sacramento area\'s weather today is ',
         .,
         ', with possible highs of ',
         highs,
         ' and possible lows of ',
         lows,
         '.')


# Post tweet
post_tweet(status = message)