scipeds 0.0.8

A Python package for working with IPEDS data

[!NOTE] 2024 data: As of v0.0.8, scipeds now includes 2024 data!

A Python package for working with IPEDS data.

Read the full documentation here.

Quickstart

Option 1: Use Colab (no installation required)

Click the link below to launch a pre-configured Google Colab notebook for scipeds:

Open the Colab notebook using the link above (also here), and then follow the instructions in the notebook to explore and use scipeds in a cloud environment. This approach does not require you to install anything on your computer.

If you want to keep using scipeds this way, you'll need to make a copy of the notebook into your own Google Drive.

Option 2: Install scipeds on your computer

Alternatively, you can install scipeds on your own computer and work from there.

Install scipeds

Open a terminal and type:

pip install scipeds

Download the pre-processed database

You can download the pre-processed database in two ways.

Either from the shell:

scipeds download-db

or from within python (i.e. in a Python interactive shell or from a notebook):

import scipeds

scipeds.download_db()

Query completions data using the corresponding query engine

Now you are ready to try scipeds's functionality!

For example, you can look at completions data by gender:

from scipeds.data.completions import CompletionsQueryEngine
from scipeds.data.queries import (
    FieldTaxonomy,
    QueryFilters, 
)

engine = CompletionsQueryEngine()

Use a pre-baked query:

gender_df = engine.field_totals_by_grouping(
    grouping="gender", 
    taxonomy=FieldTaxonomy.ncses_field_group,
    query_filters=QueryFilters()
)
gender_df.head()

or write your own using duckdb SQL syntax:

from scipeds.constants import COMPLETIONS_TABLE

df = engine.get_df_from_query(f"""
    SELECT * 
    FROM {COMPLETIONS_TABLE}
    LIMIT 10;
""")
df.head()

For more detailed usage, see the Usage page or the engine API Reference.

About scipeds

What is scipeds?

scipeds is a Python package for working with data from IPEDS. Specifically, scipeds makes it easier for people to analyze data from IPEDS by pre-processing and standardizing IPEDS data into a database and providing some Python tooling for querying that database.

scipeds is not a tool for working with raw IPEDS data. For that, you should download data directly from IPEDS.

Full scipeds documentation can be found at this link, and the source code is available on GitHub.

Currently supported IPEDS surveys

scipeds currently supports the following datasets / survey components:

• IPEDS Completions by program (6-digit CIP code), award level, race/ethnicity, and gender from 1984-2023
• IPEDS Institutional Characteristics Directory Information from 2011-2023

Completions data preprocessing

We provide functionality to reproduce our pre-processing of the IPEDS data. To recreate the pre-processed database, you can clone the scipeds repository, download the raw data, and re-run the pipeline code in pipeline/. Decisions about how to convert / crosswalk data across different years and handle other edge cases such as missing data are contained in the pipeline code.

Why does scipeds exist?

While IPEDS provides a large volume of data about higher education in the United States, working with IPEDS data can be challenging! Many things have changed in the time that data has been reported to IPEDS, making it non-trivial to join datasets across different time periods to consistently measure changes over time.

In the process of their own work, the authors found it useful to create tools to make it easier to analyze IPEDS data and hoped that the tools they created would be useful to others as well.

Who created scipeds?

scipeds was created by Science for America (in collaboration with DrivenData) as part of its mission to address urgent challenges in STEM education.

The scipeds logo was designed by Adrianna Mena.