doc_id
stringlengths
16
25
page
int64
0
635
chunk
int64
0
24
mention_name
stringlengths
2
86
start
int64
0
2.75k
end
int64
5
2.78k
extractor_specificity
stringclasses
3 values
extractor_typology
stringclasses
1 value
specificity
stringclasses
3 values
typology
stringclasses
11 values
usage
stringlengths
6
756
valid_data_mention
bool
2 classes
judge_error
stringclasses
1 value
reliefweb:001624
13
0
national census
1,444
1,459
descriptive
descriptive
other
the Government announced its intention to proceed with a ⟦national census⟧ in preparation for the national elections.
false
null
reliefweb:001624
28
0
Refugees in and from DRC
1,316
1,340
descriptive
named
other
⟦Refugees in and from DRC⟧, 1990-1999
false
null
reliefweb:001624
2
0
United Nations Population Division
345
379
named
named
estimates
The ⟦United Nations Population Division⟧ estimates the population to be 37,405,000 in 1990
true
null
reliefweb:001624
36
1
1999 Country Reports on Human Rights Practices
613
659
named
named
other
**U.S. Department of State** : ⟦1999 Country Reports on Human Rights Practices⟧, DRC, 25 February 2000
false
null
reliefweb:001625
10
0
Registered Incidents
5
25
named
descriptive
other
#### ⟦Registered Incidents⟧: October - December 2016
false
null
reliefweb:001625
11
0
Gender-Based Violence Information Management System
797
848
named
named
system
the ⟦Gender-Based Violence Information Management System⟧ (GBV-IMS), used by humanitarian service providers and managed by UNFPA, recorded 575 incidents of GBV
true
null
reliefweb:001625
11
1
GBV – IMS data
1,475
1,489
named
named
system
the ⟦GBV – IMS data⟧ cannot capture reports of service and GBV incidents
true
null
reliefweb:001625
11
2
data for December 2016
713
735
vague
vague
other
The ⟦data for December 2016⟧ is not complete yet
true
null
reliefweb:001625
12
0
Coordinated GBV and Protection Safety Audit
685
728
named
named
other
See “⟦Coordinated GBV and Protection Safety Audit⟧”, 3 to 11 November 2016
false
null
reliefweb:001625
12
0
GBVIMS
816
822
named
named
system
The ⟦GBVIMS⟧ was rolled out in South Sudan in 2014
true
null
reliefweb:001625
12
0
GBVIMS data
909
920
named
named
database
⟦GBVIMS data⟧ that is shared is only from reported cases
true
null
reliefweb:001625
12
0
GBVIMS
1,216
1,222
named
named
system
service providers who use the ⟦GBVIMS⟧ for data collection
true
null
reliefweb:001625
13
0
Mine Action database
894
914
named
named
database
the known hazards recorded in the ⟦Mine Action database⟧
true
null
reliefweb:001625
13
1
Known Hazardous Areas and Number of Known Hazards
910
959
descriptive
descriptive
other
⟦Known Hazardous Areas and Number of Known Hazards⟧ as of December 2016
false
null
reliefweb:001625
13
1
Known Hazardous Areas and Number of Known Hazards
1,015
1,064
descriptive
descriptive
other
⟦Known Hazardous Areas and Number of Known Hazards⟧ as of December 2016
false
null
reliefweb:001625
1
2
Small Arms Survey report
1,274
1,298
named
named
other
See ⟦Small Arms Survey report⟧, December 2016
false
null
reliefweb:001625
2
2
quantitative data
1,533
1,550
vague
vague
other
UNHCR teams are collecting ⟦quantitative data⟧
true
null
reliefweb:001625
2
4
DRC Intention Surveys
417
438
named
named
survey
See ⟦DRC Intention Surveys⟧ for Melut POC site and Palioch school
true
null
reliefweb:001625
2
4
MSF survey in Malakal POC
496
521
named
named
survey
and ⟦MSF survey in Malakal POC⟧ (May 2016)
true
null
reliefweb:001625
2
4
UNHCR Block Survey Report
774
799
named
named
survey
See ⟦UNHCR Block Survey Report⟧ (POC 1&3), 24 January 2017
true
null
reliefweb:001625
3
0
monitoring of new arrivals in Uganda
1,412
1,448
descriptive
descriptive
other
From ⟦monitoring of new arrivals in Uganda⟧, a number have been living as IDPs
false
null
reliefweb:001625
3
1
monitoring of new arrivals in Uganda
105
141
descriptive
descriptive
other
From ⟦monitoring of new arrivals in Uganda⟧, a number have been living as IDPs
false
null
reliefweb:001625
4
0
community assessments
815
836
descriptive
descriptive
survey
by ⟦community assessments⟧, it is estimated there are well over 100,000 IDPs in Juba
true
null
reliefweb:001625
4
1
PC Trends report
1,768
1,784
named
named
report
since the last ⟦PC Trends report⟧ at over 80,000
true
null
reliefweb:001625
4
2
PC Trends report
49
65
named
named
report
From the previous ⟦PC Trends report⟧, less incidents of attacks on civilians
true
null
reliefweb:001625
4
2
December head count
354
373
descriptive
descriptive
census
population 29,021 individuals from ⟦December head count⟧
true
null
reliefweb:001625
5
1
RRC
933
936
named
named
other
over 1000 registered with the ⟦RRC⟧
false
null
reliefweb:001625
6
0
movement tracking in the POC
280
308
descriptive
descriptive
other
⟦movement tracking in the POC⟧ identified persons returning to the site
true
null
reliefweb:001625
6
0
intention surveys
397
414
descriptive
descriptive
survey
conducting further ⟦intention surveys⟧ to establish triggers for movement
true
null
reliefweb:001625
6
1
vulnerability assessment conducted by the Protection Cluster
1,632
1,692
descriptive
descriptive
survey
A ⟦vulnerability assessment conducted by the Protection Cluster⟧ in December revealed that some 507 households
true
null
reliefweb:001625
6
2
IOM headcount
1,481
1,494
named
named
census
according to a recent ⟦IOM headcount⟧
true
null
reliefweb:001625
6
3
UNMISS –GHANBATT figures
704
728
named
named
other
according to ⟦UNMISS –GHANBATT figures⟧
true
null
reliefweb:001625
8
1
Family Tracing and Reunification database
1,360
1,401
named
named
database
Source: ⟦Family Tracing and Reunification database⟧
true
null
reliefweb:001625
9
0
Monitoring and Reporting Mechanism
557
591
named
named
system
reported through the ⟦Monitoring and Reporting Mechanism⟧ (MRM)
true
null
reliefweb:001625
9
1
Registered Incidents per State
490
520
descriptive
descriptive
other
⟦Registered Incidents per State⟧
false
null
reliefweb:001625
9
1
Monitoring and Reporting Mechanism
829
863
named
named
system
Source: ⟦Monitoring and Reporting Mechanism⟧
true
null
reliefweb:001627
1
1
Map shows civilian impact hotspots
395
429
descriptive
descriptive
geospatial
(⟦Map shows civilian impact hotspots⟧ from May 24 to June 6)
true
null
reliefweb:001627
1
2
civilian impact in Al-Hudaydah governorate
1,638
1,680
descriptive
descriptive
other
This short assessment seeks to extract analytical highlights from ⟦civilian impact in Al-Hudaydah governorate⟧[1]
false
null
reliefweb:001627
2
0
CIMP
1,057
1,061
named
named
other
the displacement recorded by ⟦CIMP⟧[2] from incidents of armed violence directly impacting on civilian homes
true
null
reliefweb:001630
12
1
Big (Crisis) Data sources
880
905
descriptive
vague
other
studies mainly used ⟦Big (Crisis) Data sources⟧[3] in manageable case studies
true
null
reliefweb:001630
13
0
CDO database
457
469
named
named
database
the World Bank’s economic indicators, NOAA’s ⟦CDO database⟧[4]
true
null
reliefweb:001630
13
0
conflict database
513
530
named
descriptive
database
ACLED’s or the university of Uppsala’s ⟦conflict database⟧[5]
true
null
reliefweb:001630
13
0
QuickBird
1,075
1,084
named
named
geospatial
Fine-resolution satellite imagery (e.g., WorldView, Pleaides, IKONOS, GEO, and ⟦QuickBird⟧[6])
true
null
reliefweb:001630
13
0
training data
1,515
1,528
vague
descriptive
other
Supervised classification methods usually generate more accurate results, but require ⟦training data⟧[7] from a considerable portion of the area to work correctly
true
null
reliefweb:001630
13
1
Omdena project
756
770
named
named
other
These two points can potentially produce bias in the data and limit the scalability of this solution (e.g., see ⟦Omdena project⟧[8])
false
null
reliefweb:001630
13
1
remote-sensing nocturnal lights images
797
835
descriptive
descriptive
geospatial
Levin et al. [22] use ⟦remote-sensing nocturnal lights images⟧[9] to track conflict areas in real-time during the Arab spring
true
null
reliefweb:001630
14
0
Global Pulse project
182
202
named
named
other
e.g., a ⟦Global Pulse project⟧[10] conducted in Sudan investigates the correlation between nocturnal lights and poverty rates
false
null
reliefweb:001630
14
0
EventRegistry
1,725
1,738
named
named
database
Projects that daily collect large quantities of media in multiple languages are, e.g., ⟦EventRegistry⟧[11]
true
null
reliefweb:001630
14
0
GDELT project
1,744
1,757
named
named
database
the ⟦GDELT project⟧[12]
true
null
reliefweb:001630
14
0
ICEWS
1,759
1,764
named
named
database
the ⟦ICEWS⟧[13]
true
null
reliefweb:001630
14
0
Expandable Open Source database
1,770
1,801
named
named
database
the ⟦Expandable Open Source database⟧[14] (EOS)
true
null
reliefweb:001630
14
0
Europe Media Monitor
1,813
1,833
named
named
database
the ⟦Europe Media Monitor⟧[15] (EMM) project
true
null
reliefweb:001630
14
0
Global Terrorist Database
1,877
1,902
named
named
database
the ⟦Global Terrorist Database⟧[16]
true
null
reliefweb:001630
14
1
Expandable Open Source database
24
55
named
named
database
the ⟦Expandable Open Source database⟧[17] (EOS)
true
null
reliefweb:001630
14
1
Europe Media Monitor
67
87
named
named
database
the ⟦Europe Media Monitor⟧[18] (EMM) project
true
null
reliefweb:001630
14
1
Global Terrorist Database
131
156
named
named
database
the ⟦Global Terrorist Database⟧[19]
true
null
reliefweb:001630
14
1
Event data
601
611
vague
descriptive
other
⟦Event data⟧[20] from news media outlets have the advantage that they are vetted before publication by the press agency
true
null
reliefweb:001630
14
1
news article databases
787
809
descriptive
descriptive
database
⟦news article databases⟧[21] allow cross-checking facts on an event through multiple reports from multiple press agencies
true
null
reliefweb:001630
15
0
news databases
931
945
descriptive
descriptive
database
Another challenge in a broader application of this ⟦news databases⟧[22] are unspecified names of localities
true
null
reliefweb:001630
15
0
GDELT database
1,690
1,704
named
named
database
Melachrinos et al. [31] use 240 different event types from the ⟦GDELT database⟧[23] to generate a push-factor-index
true
null
reliefweb:001630
15
0
Data from news media
1,775
1,795
descriptive
descriptive
other
⟦Data from news media⟧[24] are a popular source for early event detection in predictive models of forced displacement
true
null
reliefweb:001630
15
1
GDELT database
36
50
named
named
database
from the ⟦GDELT database⟧[1] to generate a push-factor-index
true
null
reliefweb:001630
15
1
EOS database
283
295
named
named
database
use event data from the ⟦EOS database⟧[2] to predict movements
true
null
reliefweb:001630
15
1
event data on violence
388
410
descriptive
descriptive
other
use ⟦event data on violence⟧[3] from the GDELT database as a proxy
true
null
reliefweb:001630
15
1
EOS database
583
595
named
named
database
collect 1.4 million English language news articles from the ⟦EOS database⟧[4]
true
null
reliefweb:001630
15
1
news article databases
712
734
descriptive
descriptive
database
use ⟦news article databases⟧[5] to develop early event detection systems as a primary data source
true
null
reliefweb:001630
15
1
data from social media sites
1,237
1,265
descriptive
descriptive
other
Although ⟦data from social media sites⟧[6] have one of the lowest signal-to-noise ratios
true
null
reliefweb:001630
15
1
data from social media sites
1,472
1,500
descriptive
descriptive
other
analysts can use ⟦data from social media sites⟧[7] in two ways
true
null
reliefweb:001630
15
1
remote-sensing data
1,646
1,665
descriptive
descriptive
geospatial
in the likeness with ⟦remote-sensing data⟧[8] of nocturnal lights
true
null
reliefweb:001630
16
0
user data
50
59
vague
descriptive
other
Because of the easy and free accessibility of its ⟦user data⟧[9], Twitter has become the most popular social media site
true
null
reliefweb:001630
16
0
data from the Twitter API
485
510
descriptive
descriptive
other
Investigation of the ⟦data from the Twitter API⟧[10] show
true
null
reliefweb:001630
16
0
GDELT database
758
772
named
named
database
combines data from the ⟦GDELT database⟧[11] with ⟦data from Twitter⟧[12]
true
null
reliefweb:001630
16
0
data from Twitter
778
795
descriptive
descriptive
other
with ⟦data from Twitter⟧[12] to develop an algorithm for early event detection
true
null
reliefweb:001630
16
1
geotagged Flickr photographs
629
657
descriptive
named
geospatial
use ⟦geotagged Flickr photographs⟧[13] as a proxy for the spread and the intensity of conflicts
true
null
reliefweb:001630
17
0
news article databases
112
134
descriptive
descriptive
database
data from ⟦news article databases⟧[14] could additionally be used to try to build prediction models
true
null
reliefweb:001630
17
0
ACLED
992
997
named
named
database
Event characteristics can be derived from either traditional data sources, e.g., ⟦ACLED⟧[15]
true
null
reliefweb:001630
17
0
Big (Crisis) Data
1,188
1,205
vague
descriptive
other
⟦Big (Crisis) Data⟧[16] can supplement with valuable and timely information
true
null
reliefweb:001630
17
0
data from internet searches or social media sites
1,486
1,535
descriptive
descriptive
other
E.g., ⟦data from internet searches or social media sites⟧[17] that inquire about flight routes or host countries
true
null
reliefweb:001630
17
0
Internet searches
1,750
1,767
descriptive
descriptive
other
**⟦Internet searches⟧[18]** ⟦Google Trends⟧[19] tracks its user’s search queries
true
null
reliefweb:001630
17
0
Google Trends
1,770
1,783
named
named
system
⟦Google Trends⟧[19] tracks its user’s search queries
true
null
reliefweb:001630
17
1
trend data
246
256
vague
descriptive
other
These ⟦trend data⟧[20] are frequently used in migration studies
true
null
reliefweb:001630
17
1
Data from Google Trends
513
536
named
named
system
⟦Data from Google Trends⟧[21] is free and easily accessible
true
null
reliefweb:001630
17
1
search data from Google Trends
817
847
named
named
system
uses ⟦search data from Google Trends⟧[22] to assess South Americans’ migration intentions
true
null
reliefweb:001630
17
1
demographic data
1,147
1,163
descriptive
descriptive
other
through ⟦demographic data⟧[23], size of the agricultural sector
true
null
reliefweb:001630
18
0
Google Trends data
62
80
named
named
system
uses ⟦Google Trends data⟧[24] to forecast migration
true
null
reliefweb:001630
18
0
trend data
1,087
1,097
vague
descriptive
other
isolate the target population in the ⟦trend data⟧[25] to generate trends
true
null
reliefweb:001630
18
0
Google Correlate
1,785
1,801
named
named
system
Data from Google Trends and ⟦Google Correlate⟧[26] are an easy to use and easy to up-scale Big (Crisis) Data source
true
null
reliefweb:001630
18
1
Google Trends data
164
182
named
named
system
The biggest weakness of ⟦Google Trends data⟧[27] is bias due to limited access to the internet
true
null
reliefweb:001630
18
1
data on internet penetration rates
239
273
descriptive
descriptive
indicator
⟦data on internet penetration rates⟧[28], which could be used to identify and possibly correct biases
true
null
reliefweb:001630
18
1
Big (Crisis) Data
651
668
descriptive
descriptive
other
⟦Big (Crisis) Data⟧[29] offers many potential data sources to capture movements and hidden populations
true
null
reliefweb:001630
18
1
call data records
836
853
descriptive
descriptive
administrative
geotagged social media posts, and ⟦call data records⟧[30] are among the most prominent examples
true
null
reliefweb:001630
19
0
time-series of QuickBird images
1,388
1,419
named
named
geospatial
used a ⟦time-series of QuickBird images⟧ to map the evolution in population size
true
null
reliefweb:001630
19
1
remote-sensing data
249
268
descriptive
descriptive
geospatial
Data gaps due to cloudiness and air pollution are the biggest disturbance in ⟦remote-sensing data⟧
true
null
reliefweb:001630
20
0
social media data
1,662
1,679
vague
descriptive
other
Studies based on ⟦social media data⟧ report mixed results in this respect
true
null
reliefweb:001630
20
0
Twitter data
1,724
1,736
named
named
database
Using ⟦Twitter data⟧, Wong et al. [47] try to solve the problem
true
null
reliefweb:001630
20
1
social media data
19
36
descriptive
descriptive
other
Studies based on ⟦social media data⟧ report mixed results in this respect
true
null
reliefweb:001630
20
1
Twitter data
81
93
named
named
database
Using ⟦Twitter data⟧, Wong et al. [47] try to solve the problem
true
null
reliefweb:001630
20
1
self-reported location data
683
710
descriptive
descriptive
other
use geotagged tweets or ⟦self-reported location data⟧ from Twitter users to estimate the number of Venezuelans
true
null
reliefweb:001630
20
1
data available through the Twitter API
903
941
descriptive
named
system
the immigration flow from Venezuela can be approximated from the ⟦data available through the Twitter API⟧
true
null
reliefweb:001630
20
1
geo-located tweets from Twitter users
1,461
1,498
descriptive
descriptive
microdata
Collecting ⟦geo-located tweets from Twitter users⟧ during the period July 2012 to June 2015 to derive their travel history
true
null