-
Notifications
You must be signed in to change notification settings - Fork 0
/
visualize_article_data.Rmd
452 lines (347 loc) · 11.1 KB
/
visualize_article_data.Rmd
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
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
---
title: "Article Metadata Summaries"
author: "Toph Allen"
date: "4/9/2019"
output: github_document
always_allow_html: yes
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = FALSE)
library(tidyverse)
library(magrittr)
library(knitr)
library(kableExtra)
library(scales)
library(here)
ktable <- function(...) {
knitr::kable(table(...)) %>%
kable_styling(bootstrap_options = c("hover", "condensed", "responsive"),
full_width = FALSE,
position = "left")
}
```
```{r load}
terms <- read_lines("terms")
articles <- read_csv(here("data", "articles.csv")) %>%
select(-X1, -id_types, -keywords)
id_types <- read_csv(here("data", "id_types.csv")) %>%
select(-X1)
keywords <- read_csv(here("data", "keywords.csv")) %>%
select(-X1)
text_matches <- read_csv(here("data", "text_matches.csv")) %>%
select(-X1)
geonames <- read_csv(here("data", "geonames.csv")) %>%
select(-X1)
subset <- read_csv(here("data", "subset.csv")) %>%
select(-X1)
```
## Prevalence of Text Matches
We searched for four terms in the extracted text of articles, using MongoDB's text search capabilities. We think these terms will select articles which have a fieldwork component over those that don't. The terms used are: `r terms`.
Matching articles are a smallish subset of the total, `r sum(articles$any_matches) / nrow(articles) * 100`%.
```{r text_matches_1}
ktable(articles$any_matches)
ktable(articles$text_matches) %>% scroll_box(height = "400px")
```
```{r text_matches_2}
text_matches %>%
filter(!is.na(text_matches)) %>%
ggplot(aes(x = factor(text_matches))) +
geom_bar(stat = "count") +
theme_bw()
```
## Examining characteristics of the subset
I pulled a few pieces of metadata for our subset of articles. These include:
- What type of article is it?
- Does it have a `<body>` tag?
- How long is it?
- Associated keywords
- What types of document identifiers does it have?
### Article types
There are many article types, but only a few which are widely used.
```{r article_types_1}
articles %<>%
mutate(article_type = fct_infreq(article_type))
ktable(articles$article_type) %>% scroll_box(height = "400px")
```
```{r article_types_2}
articles %>%
group_by(article_type) %>%
mutate(n = n()) %>%
filter(n > 50) %>%
ungroup() %>%
ggplot(aes(x = factor(article_type))) +
geom_bar() +
theme_bw() + theme(axis.text.x = element_text(angle = 45, hjust = 1))
```
Does the usage of these differ between the fieldwork subset?
```{r article_types_3}
articles %>%
group_by(any_matches) %>%
mutate(n_matches = n()) %>%
group_by(any_matches, article_type) %>%
summarize(type_freq = (n() / n_matches)[1]) %>%
ungroup() %>%
group_by(article_type) %>%
filter(n() > 1) %>%
ungroup() %>%
ggplot(aes(x = factor(article_type), y = type_freq, fill = any_matches)) +
geom_col(position = "dodge") +
theme_bw() + theme(axis.text.x = element_text(angle = 45, hjust = 1))
```
There are more research articles, and fewer of everything else.
### Presence of `<body>` tag
```{r body_tag}
ktable(articles$has_body)
ktable(articles$article_type, articles$has_body) %>% scroll_box(height = "400px")
```
### Length
```{r length_1}
types <- articles %>%
group_by(any_matches) %>%
mutate(match_n = n()) %>%
group_by(article_type) %>%
mutate(type_n = n()) %>%
group_by(any_matches, article_type) %>%
summarize(match_x_type = n(),
match_n = match_n[1],
type_n = type_n[1]) %>%
ungroup() %>%
mutate(x_over_match = match_x_type / match_n,
x_over_type = match_x_type / type_n)
type_filter <- types %>%
filter(x_over_match > 0.005) %>%
group_by(article_type) %>%
filter(n() > 1) %>%
pull(article_type) %>%
unique()
```
```{r length_2}
ggplot(articles, aes(x = text_length)) +
geom_density() +
scale_x_log10(labels = comma) +
theme_bw()
ggplot(articles, aes(x = text_length, fill = any_matches)) +
geom_density(alpha = 0.5) +
scale_x_log10(labels = comma) +
theme_bw()
articles %>%
group_by(article_type) %>%
filter(n() > 100) %>%
ggplot(aes(x = text_length, fill = article_type)) +
geom_density(alpha = 0.25) +
scale_x_log10(labels = comma) +
theme_bw()
articles %>%
filter(article_type %in% type_filter) %>%
ggplot(aes(x = text_length, fill = article_type)) +
geom_density() +
scale_x_log10(labels = comma) +
facet_grid(article_type ~ any_matches) +
theme_bw()
```
### Usage of Keywords
```{r keywords_1}
keywords_per_paper <- keywords %>%
group_by(id) %>%
summarize(any_matches = unique(any_matches),
n = n())
```
A total of `r nrow(keywords_per_paper)` papers out of `r nrow(articles)` (`r nrow(keywords_per_paper)/nrow(articles) * 100`%) had keywords.
For those matching fieldwork terms:
```{r keywords_2}
k <- keywords_per_paper %>%
filter(any_matches == TRUE) %>%
nrow()
nk <- articles %>%
filter(any_matches == TRUE) %>%
nrow()
k / nk * 100
```
For those not matching fieldwork terms:
```{r keywords_3}
k <- keywords_per_paper %>%
filter(any_matches == FALSE) %>%
nrow()
nk <- articles %>%
filter(any_matches == FALSE) %>%
nrow()
k / nk * 100
```
These are the top twenty-five keywords used in all documents:
```{r keywords_4}
keyword_summary <- keywords %>%
group_by(keywords) %>%
summarize(n = n()) %>%
arrange(-n) %>%
head(25)
knitr::kable(keyword_summary) %>%
kable_styling(bootstrap_options = c("hover", "condensed", "responsive"),
full_width = FALSE,
position = "left") %>%
scroll_box(height = "400px")
```
For those matching fieldwork terms:
```{r keywords_5}
fw_keyword_summary <- keywords %>%
filter(any_matches == TRUE) %>%
group_by(keywords) %>%
summarize(n = n()) %>%
arrange(-n) %>%
head(25)
knitr::kable(fw_keyword_summary) %>%
kable_styling(bootstrap_options = c("hover", "condensed", "responsive"),
full_width = FALSE,
position = "left") %>%
scroll_box(height = "400px")
```
And those with no matches for fieldwork terms:
```{r keywords_6}
nofw_keyword_summary <- keywords %>%
filter(any_matches == FALSE) %>%
group_by(keywords) %>%
summarize(n = n()) %>%
arrange(-n) %>%
head(25)
knitr::kable(nofw_keyword_summary) %>%
kable_styling(bootstrap_options = c("hover", "condensed", "responsive"),
full_width = FALSE,
position = "left") %>%
scroll_box(height = "400px")
```
And per fieldwork term:
```{r keywords_7}
kwd_x_match <- keywords %>%
left_join(select(articles, id, text_matches), by = "id") %>%
mutate(text_matches = replace_na(text_matches, ""))
for (term in terms) {
print(paste("Term: ", term))
kwd_x_match %>%
filter(map_lgl(kwd_x_match$text_matches, ~str_detect(.x, fixed(term)))) %>%
group_by(keywords) %>%
summarize(n = n()) %>%
arrange(desc(n)) %>%
print()
}
```
## GeoNames
It looks like articles that match the fieldwork terms have many more geonames in them than those that don't.
```{r geonames}
summary(geonames)
geonames %>% filter(any_matches == TRUE) %>% summary
geonames %>% filter(any_matches == FALSE) %>% summary
```
## Subset
We're going to take a brief look at the distribution of GeoNames and text length in the subset.
```{r subset}
summary(subset)
ggplot(subset, aes(x = length)) +
geom_histogram(binwidth = 1000) +
theme_bw()
ggplot(subset, aes(x = n_geospans)) +
geom_histogram(binwidth = 10) +
theme_bw()
```
```{r density}
ggplot(subset, aes(x = n_geospans / length)) +
geom_histogram(bins = 500) +
theme_bw()
ggplot(subset, aes(x = n_geospans / length)) +
geom_histogram(bins = 500) +
scale_x_log10(labels = comma) +
theme_bw()
```
```{r quantiles}
deciles = seq(0, 1, 0.1)
quantile(subset$length, probs = deciles)
quantile(subset$n_geospans, probs = deciles)
quantiles = c(0.05, 0.25, 0.5, 0.75, 0.95)
quantile(subset$length, probs = quantiles)
quantile(subset$n_geospans, probs = quantiles)
percentiles = seq(0, 1, 0.01)
quantile(subset$length, probs = percentiles)
quantile(subset$n_geospans, probs = percentiles)
qplot(seq(0, 1, 0.01), quantile(subset$length, probs = percentiles)) + theme_bw()
qplot(seq(0, 1, 0.01), quantile(subset$n_geospans, probs = percentiles)) + theme_bw()
```
We will only include documents which are:
- longer than 1% and shorter than 5% of documents or 60,000 characters;
- include more GeoNames than 25% of documents, and fewer than 99%.
### Density
I will also consider subsetting on `n_geospans / length`.
```{r compute_density}
subset <- subset %>%
filter(n_geospans > 0) %>%
mutate(density = n_geospans / length,
log_density = log(density))
mean(subset$log_density)
sd(subset$log_density)
subset <- subset %>%
mutate(density_exclude = ifelse(log_density < mean(log_density) - sd(log_density), 1, 0),
density_excludem = ifelse(log_density < median(log_density), 1, 0))
ggplot(subset, aes(x = log_density)) +
geom_histogram(bins = 500, aes(fill = factor(density_exclude))) +
theme_bw()
ggplot(subset, aes(x = density)) +
geom_histogram(bins = 500, aes(fill = factor(density_exclude))) +
theme_bw()
ggplot(subset, aes(x = density)) +
geom_histogram(bins = 500, aes(fill = factor(density_excludem))) +
theme_bw()
ggplot(subset, aes(x = length, y = n_geospans, color = factor(density_exclude))) +
geom_point() +
theme_bw()
ggplot(subset, aes(x = length, y = n_geospans, color = factor(density_excludem))) +
geom_point() +
theme_bw()
subset %>%
filter(density_exclude == 1) %>%
summary()
subset %>%
filter(density_exclude == 0) %>%
summary()
# Plot of density exclusion
ggplot(subset, aes(x = length, fill = factor(density_exclude))) +
geom_histogram(binwidth = 1000) +
facet_grid(density_exclude~.) +
theme_bw()
ggplot(subset, aes(x = n_geospans, fill = factor(density_exclude))) +
geom_histogram(binwidth = 10) +
facet_grid(density_exclude~.) +
theme_bw()
ggplot(subset, aes(x = length, fill = factor(density_excludem))) +
geom_histogram(binwidth = 1000) +
facet_grid(density_excludem~.) +
theme_bw()
ggplot(subset, aes(x = n_geospans, fill = factor(density_excludem))) +
geom_histogram(binwidth = 10) +
facet_grid(density_excludem~.) +
theme_bw()
# Log version
ggplot(subset, aes(x = length, fill = factor(density_exclude))) +
geom_histogram(bins = 100) +
facet_grid(density_exclude~.) +
scale_x_log10(labels = comma) +
theme_bw()
ggplot(subset, aes(x = n_geospans, fill = factor(density_exclude))) +
geom_histogram(bins = 100) +
facet_grid(density_exclude~.) +
scale_x_log10(labels = comma) +
theme_bw()
ggplot(subset, aes(x = length, fill = factor(density_excludem))) +
geom_histogram(bins = 100) +
facet_grid(density_excludem~.) +
scale_x_log10(labels = comma) +
theme_bw()
ggplot(subset, aes(x = n_geospans, fill = factor(density_excludem))) +
geom_histogram(bins = 100) +
facet_grid(density_excludem~.) +
scale_x_log10(labels = comma) +
theme_bw()
```
We will exclude articles which are more than one standard deviation below the mean of log density.
## Summary
We should use documents which
- match this set of search terms;
- have a `<body>` tag;
- are a `"research-article"`.
We could also consider including `"review-article"` and `"brief-report"` types, but those seem fairly uncommon.