-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathanimation_day_by_day.R
216 lines (167 loc) · 10.2 KB
/
animation_day_by_day.R
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
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
# David Ebert
# 1 November 2016
# Updated 7 April 2017
# Make an animation of tweet times in LA county.
# Make a bunch of .png files, then use image magick to create .gif
# Idea from https://www.r-bloggers.com/animated-plots-with-r/
library(ggmap)
library(feather)
library(gridExtra)
library(ggthemes)
#############################################################
### Import 4 weeks of data with time set
#############################################################
all_august_pred_import = read_feather(path = "~/Desktop/Huang Research/LAR_Data/all_august_pred.feather")
# shorten to a random 1 million - optional
# all_august_pred_import = all_august_pred_import[sample(1:nrow(all_august_pred_import),1000000),]
# convert to POSIXct time and fix time zone (subtract 7 hours to get to PDT time zone)
head(all_august_pred_import$created_at)
all_august_pred_import$created_at = strptime(all_august_pred_import$created_at, "%a %b %d %H:%M:%S %z %Y", tz = "GMT")
#"Mon Aug 01 05:09:35 +0000 2016"
head(all_august_pred_import$created_at)
all_august_pred_import$created_at = as.POSIXct(all_august_pred_import$created_at, tz = "GMT")
head(all_august_pred_import$created_at)
all_august_pred_import$created_at = format(all_august_pred_import$created_at, tz = "America/Los_Angeles", usetz = TRUE)
all_august_pred_import$created_at = strptime(all_august_pred_import$created_at, "%Y-%m-%d %H:%M:%S", tz = "PDT")
#"2016-07-31 22:09:35 PDT"
head(all_august_pred_import$created_at) # Time should now start at 10:09 PM on the 31st, not 5 AM on the 1st
# Note that this makes sense since we cut off tweet collection just after midnight CST = 5 AM GMT = 10 PM PDT
# Use help(POSIXlt) to see subsetting types
table(all_august_pred_import$created_at$wday)
table(all_august_pred_import$created_at$mday)
table(all_august_pred_import$created_at$mon)
# Select exactly 4 weeks of data:
all_august_pred_import = all_august_pred_import[all_august_pred_import$created_at$mon==7,] # remove a few tweets from July
all_august_pred_import = all_august_pred_import[all_august_pred_import$created_at$mday>=2,] # remove tweets from the 1st of the month
all_august_pred_import = all_august_pred_import[all_august_pred_import$created_at$mday<=29,] # remove tweets from the 30th and 31st
length(table(all_august_pred_import$created_at$mday)) #28 days of month used
table(all_august_pred_import$created_at$mday)
table(all_august_pred_import$created_at$wday)
table(all_august_pred_import$created_at$hour)
#############################################################
### Initialize ggmap
#############################################################
# Initialize ggmap called basemap
la_county <- subset(map_data("county"), region == 'california' & subregion == 'los angeles')
orange_county <- subset(map_data("county"), region == 'california' & subregion == 'orange')
la_box = c(-119.2, 33.4, -116.6, 34.9) #This is zoomed out, zoom = 8
la_box = c(-119.3, 33.4, -116.6, 34.9) #This is zoomed in, zoom = 9
#LA_map = get_map(la_box, maptype = 'roadmap', color = 'bw', source = "osm") # original version
#LA_map = get_map(la_box, maptype = 'toner-lite', source = "google")
LA_map = get_map(la_box, maptype = 'roadmap', source = "google", zoom = 8) #zoom = 9 isn't working
#LA_map = get_map(la_box, maptype = 'terrain', source = "google")
basemap = ggmap(LA_map) +
geom_polygon(data = la_county, aes(x=long, y=lat, group = group), fill = 'blue', alpha = 0.01, color = "black") +
geom_polygon(data = orange_county, aes(x=long, y=lat, group = group), fill = 'blue', alpha = 0.01, color = "black") +
xlab("Longitude") +
ylab('Latitude') +
theme(text = element_text(size=16), legend.position = 'right') +
scale_x_continuous(limits = c(-119.15, -116.6), expand = c(0,0)) +
scale_y_continuous(limits = c(33.4, 34.8), expand = c(0,0))
basemap
#############################################################
### Initialize timeline
#############################################################
# Initialize time frames...
time_frames = expand.grid(0:23,0:6,-1,-1)
colnames(time_frames) = c('hour', 'day','avg_sentiment', 'num_tweets')
time_frames$day_frac = time_frames$day + time_frames$hour/24
a = Sys.time() # about 2 mins
for(i in 1:nrow(time_frames)){
# Find hour and day for plot title
hour = paste(time_frames[i,1],':00',sep = "")
day = c("Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday")[time_frames[i,2]+1]
print(paste(hour, day))
# Get hour data
hour_data = all_august_pred_import[all_august_pred_import$created_at$wday == time_frames[i,2]
& all_august_pred_import$created_at$hour == time_frames[i,1],]
hour_data = hour_data[sample(nrow(hour_data),nrow(hour_data)),] # Scramble data order for plotting
# Collect information about average sentiment and number of tweets
time_frames$avg_sentiment[i] = mean(hour_data$pred_polarity)
time_frames$num_tweets[i] = nrow(hour_data)/4
}
time_frames$num_tweets_scaled = time_frames$num_tweets/max(time_frames$num_tweets)
Sys.time()-a
#save time_frames to file using feather
write_feather(x = time_frames, path = 'images/time_frames.feather')
# initialize ggplot timeline
timeline = ggplot(time_frames, aes(x = day_frac, y = num_tweets_scaled, color = "Number of \n tweets")) +
geom_rect(aes(xmin=1, xmax=2, ymin=0, ymax=1), color = 'grey95', fill = 'grey95') +
geom_rect(aes(xmin=3, xmax=4, ymin=0, ymax=1), color = 'grey95', fill = 'grey95') +
geom_rect(aes(xmin=5, xmax=6, ymin=0, ymax=1), color = 'grey95', fill = 'grey95') +
geom_line(size = 1) +
geom_line(data = time_frames, aes(x = day_frac, y = avg_sentiment, color = "Average \n polarity"), size = 1) +
theme_few() +
theme(axis.title.y = element_blank(),
axis.text.y = element_blank(),
axis.title.x = element_blank(),
axis.ticks.y = element_blank(),
legend.position = 'left',
legend.key.size = unit(2, 'lines'),
text = element_text(size = 16)) +
scale_x_continuous(breaks = seq(0.5,6.5,0.5),
labels = c("Sun","", "Mon","", "Tue","", "Wed","", "Thu","", "Fri","", "Sat"),
limits = c(0,7), expand = c(0,0)) +
scale_color_manual(name = "", values = c("Number of \n tweets" = 'dark green', "Average \n polarity" = 'purple'))
timeline
#############################################################
### Use ggplot to plot each hour of the week
#############################################################
# Read time_frames from feather
time_frames = read_feather(path = 'images/time_frames.feather')
time_frames = as.data.frame(time_frames)
# Initialize saturday night time_frame data frames
prev_prev_hour_data = all_august_pred_import[all_august_pred_import$created_at$wday == time_frames[nrow(time_frames)-1,2]
& all_august_pred_import$created_at$hour == time_frames[nrow(time_frames)-1,1], ]
prev_prev_hour_data = prev_prev_hour_data[sample(nrow(prev_prev_hour_data),nrow(prev_prev_hour_data)),] # Scramble data order for plotting
prev_hour_data = all_august_pred_import[all_august_pred_import$created_at$wday == time_frames[nrow(time_frames),2]
& all_august_pred_import$created_at$hour == time_frames[nrow(time_frames),1], ]
prev_hour_data = prev_hour_data[sample(nrow(prev_hour_data),nrow(prev_hour_data)),] # Scramble data order for plotting
#initialize junk tweets with polarity of 0 and 1 for purpose of ensuring color gradient
junk = all_august_pred_import[1:2,]
junk$pred_polarity = c(0,1)
# Big for loop creating plots!
a = Sys.time() # about 30-45 mins
for(i in 1:nrow(time_frames)){
# creating a name for each plot file with leading zeros
if (i < 10) {name = paste('img000',i,'.png',sep='')}
if (i < 100 && i >= 10) {name = paste('img00',i,'.png', sep='')}
if (i >= 100) {name = paste('img0', i,'.png', sep='')}
# Find hour and day for plot title
hour = paste(time_frames[i,1],':00',sep = "")
day = c("Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday")[time_frames[i,2]+1]
print(paste(hour, day))
# Get hour data
hour_data = all_august_pred_import[all_august_pred_import$created_at$wday == time_frames[i,2]
& all_august_pred_import$created_at$hour == time_frames[i,1],]
hour_data = hour_data[sample(nrow(hour_data),nrow(hour_data)),] # Scramble data order for plotting
# Collect information about average sentiment and number of tweets
time_frames$avg_sentiment[i] = mean(hour_data$pred_polarity)
time_frames$num_tweets[i] = nrow(hour_data)/4
# Map plot
plot.title = "Hourly Sentiment Polarity of Geotagged LA County Tweets"
plot.subtitle = paste(day, " ", hour, " ~ August 2-29, 2016", sep = "")
map_plot = basemap +
geom_point(data = junk, aes(x = lon, y = lat, color = pred_polarity), alpha = 0) + #ensure color gradient
geom_point(data = prev_prev_hour_data, aes(x = lon, y = lat, color = pred_polarity), alpha = 0.05) + #2 hours old
geom_point(data = prev_hour_data, aes(x = lon, y = lat, color = pred_polarity), alpha = 0.1) + #1 hour old
geom_point(data = hour_data, aes(x = lon, y = lat, color = pred_polarity), alpha = 1) + #current hour
scale_colour_gradient(low = "blue", high = "red", breaks = c(0,.5,1),
labels = c("negative", "neutral", "positive"),
guide = guide_legend(title = NULL, reverse = TRUE)) +
ggtitle(bquote(atop(.(plot.title), atop(italic(.(plot.subtitle)), ""))))
# Timeline plot
timeline_plot = timeline + geom_vline(xintercept = time_frames$day_frac[i], color = 'black', size = 2)
# Combine and save plots
combined_plot = grid.arrange(map_plot, timeline_plot, ncol = 1, heights = c(7,1))
combined_plot
ggsave(plot = combined_plot, name, path = './images', width = 14, height = 10, units = 'in')
#Move data back an hour
prev_prev_hour_data = prev_hour_data
prev_hour_data = hour_data
}
Sys.time()-a
################################################
### Bash code for building animation with FFmpeg:
################################################
#ffmpeg -framerate 4 -i img%04d.png -c:v libx264 -r 30 -pix_fmt yuv420p out.mp4