Replication Data for: CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media (doi:10.7910/DVN/SS4LNN)

View:

Part 1: Document Description
Part 2: Study Description
Part 3: Data Files Description
Part 4: Variable Description
Part 5: Other Study-Related Materials
Entire Codebook

Replication Data for: CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media

doi:10.7910/DVN/SS4LNN

Harvard Dataverse

2019-06-25

1

Pan, Jennifer; Zhang, Han, 2019, "Replication Data for: CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media", https://doi.org/10.7910/DVN/SS4LNN, Harvard Dataverse, V1, UNF:6:26KE5u8/rqgZAoNS8X8wXg== [fileUNF]

Replication Data for: CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media

doi:10.7910/DVN/SS4LNN

Pan, Jennifer (Stanford University, Department of Communication)

Zhang, Han (Princeton University, Department of Sociology)

Harvard Dataverse

Pan, Jennifer

Zhang, Pan

Pan, Jennifer

2019-06-25

https://doi.org/10.7910/DVN/SS4LNN

Social Sciences, collective action, deep learning, event data, social media, China

Protest event analysis is an important method for the study of collective action and social movements and typically draws on traditional media reports as the data source. We introduce collective action from social media (CASM)—a system that uses convolutional neural networks on image data and recurrent neural networks with long short-term memory on text data in a two-stage classifier to identify social media posts about offline collective action. We implement CASM on Chinese social media data and identify more than 100,000 collective action events from 2010 to 2017 (CASM-China). We evaluate the performance of CASM through cross-validation, out-of-sample validation, and comparisons with other protest data sets. We assess the effect of online censorship and find it does not substantially limit our identification of events. Compared to other protest data sets, CASM-China identifies relatively more rural, land-related protests and relatively few collective action events related to ethnic and religious conflict.

We recommend you view files in Tree structure and begin with the readme.txt

<a href="http://creativecommons.org/publicdomain/zero/1.0">CC0 1.0</a>

Zhang Han and Jennifer Pan. 2019. “CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media” Sociological Methodology 49: 1-59.

10.1177/0081175019860244

Zhang Han and Jennifer Pan. 2019. “CASM: A Deep-Learning Approach for Identifying Collective Action Events with Text and Image Data from Social Media” Sociological Methodology 49: 1-59.

• Number of cases: 136330

• No. of variables per record: 1

• Type of File: text/tab-separated-values

UNF:6:wbuMQn7wm4tl11Io3al6oQ==

• Number of cases: 180

• No. of variables per record: 1

• Type of File: text/tab-separated-values

UNF:6:ZxN0FDr+XlhoWkcNZctu9w==

Also used for plot/SuppFigure7

• Number of cases: 6046

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:hxAPly5qKIGA1G6HfLHv1Q==

• Number of cases: 2408

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:rL3sAGsTp2RrS1yDdkgrcQ==

• Number of cases: 6413

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:mkyvOgenEhyzR7efni3q8Q==

• Number of cases: 3853

• No. of variables per record: 4

• Type of File: text/tab-separated-values

UNF:6:NKAiaoqeTkNJgn5gbpy7Cg==

• Number of cases: 4500

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:BOr7j3rkVv5QTcQtfK23Tw==

• Number of cases: 2560

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:qHEFiyMlqC8j0XwBEChMJQ==

• Number of cases: 6038

• No. of variables per record: 3

• Type of File: text/tab-separated-values

UNF:6:VueXQuSkwhSwntSBqd6gWg==

• event_id forms issues - event_id forms issues
• month count_all month_numeric quarter Dataset year - month count_all month_numeric quarter Dataset year
• label - label
• precision - precision
• recall - recall
• label - label
• precision - precision
• recall - recall
• label - label
• precision - precision
• recall - recall
• label - label
• precision - precision
• recall - recall
• threshold - threshold
• month - month
• nunique - nunique
• rank - rank
• precision - precision
• rank - rank
• recall - recall
• label - label
• precision - precision
• recall - recall

f3455389 Location:

Variable Format: character

Notes: UNF:6:wbuMQn7wm4tl11Io3al6oQ==

f3455432 Location:

Variable Format: character

Notes: UNF:6:ZxN0FDr+XlhoWkcNZctu9w==

f3455413 Location:

Variable Format: character

Notes: UNF:6:/WKB02PwmgYsUaAvUUl4WA==

f3455413 Location:

Summary Statistics: Valid 6046.0; Mean 0.3928149639659638; Min. 0.16902784562774792; Max. 1.0; StDev 0.21272211487785242

Variable Format: numeric

Notes: UNF:6:HMKRgUbpYsQmCO9gH/xqng==

f3455413 Location:

Summary Statistics: Max. 1.0; StDev 0.24434564628902133; Mean 0.8316877924352603; Min. 0.002890173410404624; Valid 6046.0

Variable Format: numeric

Notes: UNF:6:XkzU2089TYovplWPOCq1sg==

f3455404 Location:

Variable Format: character

Notes: UNF:6:VbmLr5BCRwmU99POlnD+WQ==

f3455404 Location:

Summary Statistics: Valid 2408.0; StDev 0.19843002381807667; Mean 0.6561784132954082; Min. 0.3862520458265139; Max. 1.0

Variable Format: numeric

Notes: UNF:6:ePL9lu2FbYZ9gjX7w8j01A==

f3455404 Location:

Summary Statistics: Min. 0.00211864406779661; Valid 2408.0; Max. 1.0; StDev 0.2784397034783831; Mean 0.7100924179289374

Variable Format: numeric

Notes: UNF:6:uiEp/74k1CTmxZPLsd+HmA==

f3455403 Location:

Variable Format: character

Notes: UNF:6:fPFJY05MN31qzjjPmzkzpA==

f3455403 Location:

Summary Statistics: Valid 6413.0; Min. 0.35019241341396373; StDev 0.2056319379073097; Mean 0.8141775929693685; Max. 1.0

Variable Format: numeric

Notes: UNF:6:bBv3mtaY3MeaqBDnvSY4VA==

f3455403 Location:

Summary Statistics: Max. 0.9996659242761693; StDev 0.3165098962640183; Min. 0.0015082956259426848; Mean 0.6208728005956752; Valid 6413.0

Variable Format: numeric

Notes: UNF:6:K8kuyfoC2yYRo65HurFLlg==

f3455402 Location:

Variable Format: character

Notes: UNF:6:RvEjpk9rT+6KKdRzq5zt9Q==

f3455402 Location:

Summary Statistics: Max. 1.0; Mean 0.6848123751991867; StDev 0.19064082014163491; Valid 3853.0; Min. 0.3965665236051502;

Variable Format: numeric

Notes: UNF:6:c3Q63cuB0AIvGzLYCeTvOg==

f3455402 Location:

Summary Statistics: Max. 1.0; Mean 0.7029771444704974; StDev 0.28192710238756225; Min. 0.0021598272138228943; Valid 3853.0

Variable Format: numeric

Notes: UNF:6:4Fcvgiujxg8/SJElDdlYzg==

f3455402 Location:

Summary Statistics: Max. 1.0919091868846813; Mean 0.7705188097591968; Valid 3853.0; StDev 0.20437443582886783; Min. 0.16134203970432281

Variable Format: numeric

Notes: UNF:6:9T2CstOvsCNfzAarpKYidA==

f3455410 Location:

Variable Format: character

Notes: UNF:6:O/XMkrySXj7/stLtO+oGrw==

f3455410 Location:

Summary Statistics: Max. 3126.0; Mean 1306.645555555568; Valid 4500.0; Min. 2.0; StDev 775.2087042562923

Variable Format: numeric

Notes: UNF:6:IRufL0DUEM+jNmdzQo6MJg==

f3455410 Location:

Summary Statistics: Valid 4500.0; Min. 1.0; StDev 14.432473386915156; Mean 25.5; Max. 50.0;

Variable Format: numeric

Notes: UNF:6:ikVCsyzjacSJi7zCXrf62Q==

f3455426 Location:

Summary Statistics: Mean 0.8924112448859461; StDev 0.1222086015892413; Max. 0.9831081081081081; Min. 0.24090203404994343; Valid 2560.0

Variable Format: numeric

Notes: UNF:6:5yicMrkeNYov1JxVvD9UEw==

f3455426 Location:

Summary Statistics: Min. 10.0; StDev 14.14489856907796; Valid 2560.0; Max. 50.0; Mean 30.0

Variable Format: numeric

Notes: UNF:6:19Zex4EAgUSxVumQ8R8b6w==

f3455426 Location:

Summary Statistics: Max. 1.0; StDev 0.21225415939853337; Min. 0.03178590933915893; Valid 2560.0; Mean 0.622310502768809;

Variable Format: numeric

Notes: UNF:6:rYMxh1EOHDUGy5TlCF94cQ==

f3455429 Location:

Variable Format: character

Notes: UNF:6:TWro7UTlIZoER+ZhoImqnw==

f3455429 Location:

Summary Statistics: Min. 0.0; Max. 1.0; Mean 0.3698565312123263; Valid 6038.0; StDev 0.16754473718512444

Variable Format: numeric

Notes: UNF:6:tBzZODFBTMf0oJ7mvsPsSw==

f3455429 Location:

Summary Statistics: Max. 1.0; Min. 0.0; Mean 0.8309617126216056; StDev 0.26159205941475533; Valid 6038.0;

Variable Format: numeric

Notes: UNF:6:I3gaMv8h51OylDenGlcv2g==

Label:

readme.txt

text/plain

Label:

requirements.txt

text/plain

Label:

CASM_c1_deep_text.py

text/x-python-script

Label:

CASM_c2_deep_text.py

text/x-python-script

Label:

CASM_generate_predicted_probability_image.py

text/x-python-script

Label:

CASM_generate_predicted_probability_text.py

text/x-python-script

Label:

common_operations.py

text/x-python-script

Label:

dependency.py

text/x-python-script

Label:

LSTM_text_dependency.py

text/x-python-script

Label:

word_preprocessing.py

text/x-python-script

Label:

image.json

application/json

Label:

text-stage1.json

application/json

Label:

text-stage2.json

application/json

Label:

weights_image.h5

application/x-h5

Label:

weights_text-stage1.h5

application/x-h5

Label:

weights_text-stage2.hdf5

application/x-hdf5

Label:

protest_posts.csv

text/csv

Label:

protest_region_count_prefecture_log_scale.R

type/x-r-syntax

Label:

CN-shi-A.dbf

application/dbf

Label:

CN-shi-A.prj

application/prj

Label:

CN-shi-A.sbn

application/sbn

Label:

CN-shi-A.sbx

application/sbx

Label:

CN-shi-A.shp

application/shp

Label:

CN-shi-A.shx

application/shx

Label:

CASM_divide_irrelevant.R

type/x-r-syntax

Label:

CASM_Wickedonna.R

type/x-r-syntax

Label:

wicked_events_2018-11-29.csv

text/csv

Label:

firststage_c1_text_image_precision_recall.pdf

application/pdf

Label:

firststage_c1_text_image_precision_recall.r

type/x-r-syntax

Label:

c1c2_vs_c1_precision_recall.pdf

application/pdf

Label:

c1c2_vs_c1_precision_recall.r

type/x-r-syntax

Label:

c1c2_vs_cross_validation_precision_recall.pdf

application/pdf

Label:

c1c2_vs_cross_validation_precision_recall.r

type/x-r-syntax

Label:

keywordCoverageSet.txt

text/plain

Label:

keyword_coverage_wickedonna.pdf

application/pdf

Label:

supfig1.r

type/x-r-syntax

Label:

keywordCoverageSet_keyword_year.txt

text/plain

Label:

keyword_coverage_wickedonna_by_year.pdf

application/pdf

Label:

supfig2.r

type/x-r-syntax

Label:

keywordCoverageSet_keyword_prov.txt

text/plain

Label:

keyword_coverage_wickedonna_by_province.pdf

application/pdf

Label:

supfig3.r

type/x-r-syntax

Label:

keyword_vs_num_event_700000.pdf

application/pdf

Label:

supfig4.r

type/x-r-syntax

Label:

precision_vs_keyword.pdf

application/pdf

Label:

supfig5.r

type/x-r-syntax

Label:

firststage_c1_SVM_NB_precision_recall.pdf

application/pdf

Label:

supfig6.r

type/x-r-syntax

Label:

CASM_keyword_irrelevant_count.pdf

application/pdf

Label:

supfig7.R

type/x-r-syntax

Label:

high_frequency_protest_words.txt

text/plain

Label:

jieba.dict.big.txt

text/plain

Label:

propaganda_media_words.txt

text/plain

Label:

stopwords1.txt

text/plain

Label:

village_level_dict.txt

text/plain

Label:

village_level_dict_reversed.txt

text/plain

Label:

vocab_pos_grievance.dict

application/octet-stream

Label:

vocab_pos_KGP_50000.dict

application/octet-stream