hw_text_dataset / README.md
ddecosmo's picture
Update README.md
c7e1ae8 verified
metadata
license: mit
language:
  - en
pretty_name: HW1 Text Dataset

Dataset Card for {{ pretty_name | default("Dataset Name", true) }}

This dataset covers 1000 locations at 5 universities with descriptions and reviews of locations based on my personal experience.

Dataset Details

Dataset Description

  • Curated by: Carnegie Mellon University: 24-679
  • Shared by [optional]: Devin DeCosmo
  • Language(s) (NLP): English
  • License: MIT

Dataset Sources [optional]

  • Repository: {{ repo | default("[More Information Needed]", true)}}

Uses

The main use was to train text models to predict the specific university or location based on the personal text description.

Direct Use

The direct use would be analysis tasks to predict the location based off the text.

Out-of-Scope Use

Could be used for sentiment analysis and similar tasks.

Dataset Structure

This dataset consists of two splits An original split with 100 rows An artificial split with 1000 rows

The tasks fall into 4 categories based on the university

  1. CMU
  2. ASU
  3. Bucknell
  4. Duquesne
  5. UCSD

Dataset Creation

Source Data

Source data is personal statements and reviews from the owner.

Data Collection and Processing

Data was written by the creator of this dataset.

Who are the source data producers?

Data was initially produced by the owner.

Bias, Risks, and Limitations

This is a very small data set and will likely have issues with training and fitting, especially for more complex analysis.

Recommendations

This dataset probably has limited accuracy as a first draft but may be useful for learning how to train models.