Dark Mode
Historical Spelling Variation Dataset
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This collection of names and associated information is designed for exploring and implementing phonetic hashing algorithms. It contains over 1000 names and their gender details, allowing users to predict gender based on the phonics of a name. The dataset serves as a foundational resource for solving the unique challenge of searching for names in a database when the spelling may not be accurate, a situation often caused by cultural differences, transcription errors, or a general lack of standardized spelling. Developers can use this resource to compute hash values in advance, enabling comparison of words based on how they sound, rather than their precise textual spelling.
Columns
The file
babynames_nysiis.csv contains the following fields:- babynysiis: The names themselves, consisting of 17,140 unique values, which are 100% valid.
- perc_female: The calculated percentage probability that a given name is female. (Note: Current data indicates these fields are entirely missing.)
- perc_male: The calculated percentage probability that a given name is male. (Note: Current data indicates these fields are entirely missing.)
Distribution
The data is usually provided in CSV format, containing 17,140 total values across three columns. The core content includes names and corresponding gender information. Although the
babynysiis column is fully validated, the percentage probability columns for gender (female and male) currently show that all 17.1k values are missing. The data is static and is not expected to receive future updates.Usage
- Developing and testing phonetic hash algorithms, such as the implementation of Soundex.
- Building models to predict gender likelihood solely based on the phonetic structure of a name.
- Creating robust fuzzy matching and string-search solutions for large, historical datasets where spelling errors are prevalent.
- Academic research into how spelling discrepancies affect data retrieval in genealogical and historical contexts.
Coverage
The dataset includes a variety of names (17,140 records). While specific geographic or time frame details are not explicitly noted, the inspiration for its creation relates heavily to historical records, such as those associated with the U.S. Census, which often suffer from transcription challenges. The focus is demographic, relating names to gender probability.
License
CC0: Public Domain
Who Can Use It
- Genealogists and Historians: Implementing robust search filters to overcome common spelling errors found in historical records.
- Data Engineers: Creating efficient pre-computation methods (hashing) for name matching across large data lakes.
- Data Scientists: Experimenting with fuzzy libraries in Python and exploring the predictive power of name phonics.
- Students/Educators: Learning about early efforts in phonetic encoding like the Soundex algorithm.
Dataset Name Suggestions
- Phonics-based Gender Prediction Data
- Name Soundex Algorithm Test Set
- Historical Spelling Variation Dataset
- Baby Names Phonetic Matcher
Attributes
Original Data Source: Historical Spelling Variation Dataset
QUALITYLoading...