finnsurveytext 2.1.1

• fix bug which was counting number of respondents to the open-ended question incorrectly. (Previously it was the number of overall survey respondents.)

finnsurveytext 2.1.0

• New function fst_print_available_models() to print models available through udpipe package
• Bug fix in fst_find_stopwords - remove hardcoded language of fi.
• Bug fix in fst_pos to set NAs in count to 0
• New English sample data

finnsurveytext 2.0.0

01_prepare and 01b_prepare_svydesign

• fst_format() renamed from fst_format_conllu().
• fst_prepare() renamed from fst_prepare_conllu().
• In fst_format() and fst_prepare() new parameters weights = NULL and add_cols = NULL, and parameter field renamed question
• fst_rm_stop_punct() now enables user-input list of stopwords.
• fst_format_svydesign() and fst_prepare_svydesign() introduced.

02_data_exploration

• fst_freq_table() renamed from fst_get_top_words() to reflect naming conventions.
• fst_ngrams_table() renamed from fst_get_top_ngrams() to reflect naming conventions.
• Within fst_freq_table(), fst_ngrams_table(), fst_freq(), fst_ngrams(), and fst_wordcloud() parameters to enable weights from svydesign object or from additional weights column in formatted data (use_svydesign_weights = FALSE, id = ““, svydesign = NULL, use_column_weights = FALSE).
• pos_filter parameter now accepts strings (in addition to lists of strings).

03_concept_network

• (No major changes.)

04_comparison_functions

• Data is now split into groups from one dataset (data) via a parameter called ‘field’ which is the column in formatted data that splits into groups. Each different value in this column is used to create a group which is named after the value.
• fst_freq_compare(), fst_ngrams_compare(), fst_pos_compare(), fst_summarise_compare(), fst_length_compare() and fst_comparison_cloud() now have parameters to enable inclusion/exclusion of null values when splitting data into groups (exclude_nulls = FALSE, rename_nulls = ‘null_data’)
• Default for ‘norm’ parameter is now NULL to support use of weights.
• fst_comparison_cloud() created.
• fst_freq_compare(), fst_ngrams_compare() and fst_comparison_cloud() now have parameters to enable weights from svydesign object or from additional weights column in formatted data (use_svydesign_weights = FALSE, id = ““, svydesign = NULL, use_column_weights = FALSE).
• fst_ngrams_compare_plot() has new parameter title_size = 20.
• fst_freq_compare(), and fst_ngrams_compare() have new parameters title_size = 20 and subtitle_size = 15.

05_comparison_concept_network

• As above, data split by parameter ‘field’, parameters for inclusion/exclusion of null values in ‘field’, default norm = NULL, title_size and subtitle_size parameters.
• fst_cn_get_unique() renamed fst_cn_get_unique _separate() and new function fst_cn_get_unique() created which finds unique ngrams in a list of ngrams tables.

Data

• New simplified sample data.
• “child”: Open-ended questions and background variables for Child Barometer survey.
• “dev_coop*: Open-ended questions and background variables for Development Cooperation survey.
• fst_child: Formatted data with b/g variables and weights.
• fst_child_2: Formatted data without b/g variables and weights.
• fst_dev_coop: Formatted data with b/g variables and weights.
• fst_dev_coop_2: Formatted data without b/g variables and weights.
• Removed data: child_barometer_data, conllu_cb_bullying, conllu_cb_bullying_iso, conllu_dev_q11_1_f_nltk, conllu_dev_q11_1_f, conllu_dev_q11_1_m_nltk, conllu_dev_q11_1_m, conllu_dev_q11_1_na_nltk, conllu_dev_q11_1_na, conllu_dev_q11_1_nltk, conllu_dev_q11_1, conllu_dev_q11_2_nltk, conllu_dev_q11_2, conllu_dev_q11_3_nltk, conllu_dev_q11_3, dev_data_f, dev_data_m, dev_data_na, dev_data.

finnsurveytext 1.0.0

finnsurveytext 0.1.0

• Added a NEWS.md file to track changes to the package.