• The application is supporting test with nose
• main.py is returning names in english and spanish
• api2gender.py is for return names from main apis
• confusion.py is giving support for main apis
• accuracy.py is giving support for main apis
• errors.py is giving support for genderize, damegender, genderguesser and nameapi
• apikeyadd.py allows create a password file for apis
• csv2gender.py is only for partial.csv and all.csv
• damemodels.py is creating sav files for machine learning algorithms with scikit
• gendergoogle.py is a prototype to calculate gender from google results with a name
• git2gender.py is a prototype to return number of males and females in a git repository
• mail2gender.py is a prototype to return number of males and females in a mailing list
• infofeatures.py is related to last_letter_a, last_letter_consonant, last_letter_vocal
• pca support is only a prototype

• Improved the test system with testing from python commands with bash.
• Added nameincountries.py to guess a name in different countries
• main.py: added bernoulliNB ML algorithm and support to genderguesser
• confusion.py: adding support to different dimensions
• accuracy.py: added bernoulliNB ML algorithm
• errors.py: adding genderapi and namsor support
• csv2gender.py: Minor chances
• rewriting damemodels.py to postinstall.py to recreate all files created with scripts from original files (not only ML models)
• mail2gender.py: small fix, avoid duplicated
• infofeatures.py: is related to letter_a, last_letter_a, last_letter_o, last_letter_consonant, last_letter_vocal, first_letter, first_letter_consonant, first_letter_vocal
• pca support with pca-components.py and pca-features.py

• Added downloadjson.py to download names from csv to one json file we have rewrited accuracy.py, confusion.py and errors.py to make this calculus offline

• Added damegender2json.py to generate files with ML results.
• Created ML json files
• Added new ML algorithms: tree and mlp (neural network)

• [X] United Kingdom
• [X] United States of America
• [X] Uruguay
• [X] Lucía Santamaría and Helena

• Recreated all datamodels with new datasets. Augmented the accuracies with this feature!
• Namsor stuff has been updated to Namsor2

• roc.py: to deploy roc curves to measure ML
• regenerate-ml-json.sh: execute all options of damegender2json.py to generate all json needed
• logs-accuracies.sh: execute all options of accuracy.py to generate all logs needed
• logs-confusion.sh: execute all options of confusion.py to generate all logs needed
• api2gender.py: starting the option to use wikidata

• count-debian-gender.py: add example to count males and females in debian keyring
• ethnicity.py: race about a name. Source: USA census
• surname.py: guess surname. Spain and United States of America supported
• surnameincountries.py: about countries where a surname appears. Source: INE
• readme.sh: now you can convert the readme from org to markdon with this script
• add adaboost ml algorithm

• Canada (107339 names)
• New Zealand (6600 names)
• Australia (52978 names)
• Portugal (3999 names)
• logs-errors.sh: execute allnoundefined.csv with different ML models of errors.py to generate all logs needed
• errors.py: refactorize source
• articles/damegender.pdf: updating
• files/datamodels/*sav: updtating
• manual/damegender.pdf: creating manual

• Ireland (382 names)
• Iceland (326 names)
• Finland (11449 names)
• manual/damegender.*: some advances

• top.py: print lists about the most used names in different countries
• count-scientifics.py: counting scientifics in Spain

• count-debian-gender.py: Debian Project (https://www.debian.org)
• count-forbes.py: Forbes list about most richest people (https://www.forbes.com/)
• count-gnu.py: GNU Project (https://www.gnu.org)
• count-kernel.py: Kernel Project (https://www.kernel.org/)
• count-scientifics.py: Scientifics in Spain (https://www.webometrics.info/en/GoogleScholar/Spain)
• csv2gender.py: improved for any dataset, not only the Lucia and Helena dataset.
• git2gender.py: we have fixed a lot of bugs
• surname.py: we have added --spanish_provinces using top 50 surnames per province datasets
• divide testsbycommandsextranet in testsbycommandsperceval and testsbycommandsextraapis

• top.py: add --position argument, fix --less
• mail2gender.py and git2gender.py: it shows males and females
• app/dame_statistics.py: created due to refactor in app/dame_gender.py
• csv2gender.py: add --noshow argument
• jokes.py: created. It's about damegender tips written as jokes.
• manual: new sections
• Now we have a dual license and I add scripts to change licenses

• csv2gender.py: add --verbose argument

New names from official statistics:

• Belgium (14208 names)
• Slovenia (8788 names)
• Austria (1899 names)
• Deutchsland (22368 names)
• Mexico (16122 names)

• csv2gender.py: adding new argument

Creating inter names from all countries with Open Data:

• 259395 males
• 279863 females

New names from official Open Data statisticis

• Denmark (62072 males and 79235 females)
• France (16660 males and 19783 females)

New names:

• China (2614 females and 2614 males)
• Turkey (116114 females and 67309 males)

New commands:

• mergeinterfiles.py: merge dataset files

Refactoring and bug fixing: All source is reaching PEP8 coding style now.

Improvements in commands:

api2gender.py: you can scrap the gender from wikipedia with --api=wikipedia. Before, you can use --api=wikidata, but this way (sparql) is worst in the current state of the development.

I have written "Damegender: Towards an International and Free Dataset about Name, Gender and Frequency" comments are welcome. We have presented this paper in VI International Congress of Young Researches with a Gender Perspective. You can watch the video

You can enjoy api restful power with a search engine in the website

New command: csv2jsonapirest.py to download names with json api files

New country: Switzerland (60693 names)

Updates in commands about datasets.

New country: Sweden (1079 males and 974 females)

New country: Norway (1015 females 896 males)

New country: Russia (1434 females and 3357 males)

newspaper2gender.py: to guess the author's gender from a newspaper

Cleaning code: to reach PEP8, removing unused libraries, ...

percentage2names.py: to create datasets with a specific range of percentage of males or females

launchpad2gender.py: start script to use the perceval feature about launchpad (experimental in Damegender, now)

Improvements towards reproducible science: scripting, updates in data, ...

New features:

• get-wikidata-names.py: to download names from wikidata in a specific country
• get-wikidata-surnames.py: to download surnames from wikidata in a specific country

These wikidata datasets are a subset of datasets retrieved from statistical institutions due to that in Wikipedia/Wikidata there are important people for each country, but you can find it useful if you have not names in a country or if you need a test dataset about names and gender in a country. But you must not trust in the top, frequencies, ... given by these datasets, there are exploratory data, only.

Datasets:

• You can access to Russia names in cyrillic alphabet (ru_ru) and latin alphabet (ru_en)

Improvements towards reproducible science: scripting, updates in data, ...

Official Datasets:

• [X] Argentina: 42800 names, 24587 females, 18456 males, 77583 surnames
• [X] Great Britain: 459848 names, 34326 males, 51195 females

Great Britain has several statistical offices divided by Great Britain regions. So we must do more long calculus if you find bugs, please send to Damegender, for example, with a github issue. Thanks in advance.

Cleaning code:

• [X] Replaced lxml by xml python native, lxml was doing problems in MacOS
• [X] Reducing lines of code in top.py

Dumping names and surnames for all countries from Wikidata. surnames.py and surnamesincountries.py has been updated to give support to Wikidata.

For each country, we are choosing official statistics, but if there not official statistics from now we are allowing to the users the dataset provided by wikidata. The user will receive a messsaga about it, due to that Wikidata could be giving very few names in some countries, so the quality is worst with Wikidata.

Added the 58 most popular italian names (29 males and 29 females) retrieved from istat (the italian statistics office).

In the goal to make very easy replicate the building of the dataset, we are starting to rewrite all scripts about process orig files from official statistics from states to damegender csv files. As soon as possible the only command will be orig2.py simplifying the processing. Some datasets has been updated in the process.

The method pretty_gg_list has been rewritten allowing some combination of csv or json files as arguments. This will be very useful for a Damegender non dependent of Lucia&Helena dataset. So, very good datasets saved in files/names/names_tests will can be used for check benchmarks of accuracies, precision, f1score, recall, ...

With identical goal in mind, many methods has been added to dame_statistics.py respecting the jargon math (true_negative, false_negative, false_positive, true_positive). Remember that there are many datasets stored in files/names/names_tests and tests are welcome!

More countries has been added to orig2.py

Bug fixes in accuracy.py

• accuracy.py: bug fixes
• main.py: adding --force_whitespaces
• csv2gender.py: adding --force_whitespaces
• adding Argentina dataset
• downloadjson.py: adding --outjson
• downloadjson.py: adding --name_position
• get-wikidata-names.py: adding --outcsv argument
• get-wikidata-surnames.py: adding --outcsv argument
• mergeinterfiles.py: adding --malefemale_onlygender argument
• nameincountries.py: adding message about dataset used

• accuracy.py: bug fixes
• main.py: adding --force_whitespaces
• csv2gender.py: adding --force_whitespaces
• adding Argentina dataset
• downloadjson.py: adding --outjson
• downloadjson.py: adding --name_position
• get-wikidata-names.py: adding --outcsv argument
• get-wikidata-surnames.py: adding --outcsv argument
• mergeinterfiles.py: adding --malefemale_onlygender argument
• nameincountries.py: adding message about dataset used

• Adding Chile: 7024 females, 5433 males
• top.py: adding --outjson giving this wonderful feature: to export top names to websites
• sh/languages: adding arabic.sh, german.sh and portuguese.sh
• files/names/names_tests: adding wta-players.csv, names_leagues.csv
• Luz has been reviewing that each method has a comment explaining the method

Do you want to become a hacker? Do you want to work with a lot of females around? You can take a look to statistics to understand incompatible thoughts (added statistics about males and females of developers of interpreters)

New script!

• checkcsv.py: to check possible problems in a csv file, such as, empty rows, initials, or emails

The accuracy average of INTER dataset continues augmenting! Now is 0.92. More about it in https://easychair.org/publications/preprint/Klv8

Brazil is another open data country!

• [X] dame_brazilapi.py: has been implemented
• [X] test_dame_brazilapi.py: has been implemented
• [X] api2gender.py: updated
• [X] downloadjson.py: updated
• [X] Brazil: 1322 males, 1539 females. We want download more names in a near future
• [X] downloadcsv.py: has been created this script to download Brazil, but it will be replaced by source code in orig2.py in a near future.

Do you want to build an indicator about knowledge for United Nations taking into account women and men? DameGender could help you New stats about knowledge in the DameGender website

News in commands:

• downloadcsv.py: added genderize and genderapi
• mergeinterfiles.py: adding --nosum argument
• checkcsv.py: to enable counting long string
• get-wikidata-surnames.py: to avoid to download url as surnames

The standards has been implemented:

• RFC6350
• ISOIEC5218
• DameGender (a standard de facto)

More about the standards

• https://www.iso.org/standard/81682.html
• https://en.wikipedia.org/wiki/ISO/IEC_5218
• https://datatracker.ietf.org/doc/html/rfc6350#section-6.2.7

A new paper (spanish) has been published: "Un Dataset Internacional Acerca de Nombres, Género y Frecuencias en Damegender"

La igualdad de género es el quinto objetivo de desarrollo sostenible (ODS) para Naciones Unidas. Esta igualdad puede ser lograda midiendo, analizando datos y, creando buenas políticas con los resultados. Muchos estudios de género cuentan hombres y mujeres para explicar la posible desigualdad, por ejemplo, artículos de investigación, puestos de trabajo, calles, etc. El método tradicional de investigación es usar APIs comerciales con datos propietarios sin idea acerca de cómo los datos fueron recogidos. Los datos pueden también ser recogidos desde Wikipedia, estudios lingüísticos, sitios científicos, u oficinas estadísticas. Este enfoque está basado en recoger Datasets Abiertos (Open Datasets) que incluyen nombre, género y frecuencia desde muchas instituciones estadísticas. Así́, las tareas abordadas están basadas en unificar formatos, procesar datos y, crear pruebas para medir la precisión de los nuevos datasets. El dataset usado cubre más de 20 países en el mundo occidental trayendo miles de nombres con una precisión de acierto mayor del 90%. Esto permitirá medir brecha de género a estudiantes y académicos interesados en el fenómeno sin costes y de una manera reproducible y más personas estarán contribuyendo a eliminar la brecha de género. El Software Libre y los datos provistos por instituciones estadísticas hacen posible producir investigación reproducible por pares.