QGuider

QGuider is a Python library for downloading, parsing, and querying Harvard QGuides — the university's course evaluation reports. It scrapes QGuide HTML pages, normalizes the data into typed Pydantic models, and provides a fluent API for filtering and exporting results.

Features

• Download QGuides for multiple semesters with checkpointing and resume support
• Parse HTML reports into structured, typed models
• Filter by semester, subject, department, or instructor
• Aggregate multi-instructor courses into a single record
• Export to JSON or pandas DataFrame
• Import previously exported JSON back into model objects

Installation

pip install qguider

Or install from source:

git clone https://github.com/ivanharvard/qguider
cd qguider
pip install .

Setup

QGuider requires your Harvard Key credentials to access the QGuide portal.

1. Login to Harvard QGuide Portal.
2. Either (press F12 on your keyboard) or (right click anywhere on the page, and click Inspect).
3. Press the arrow pointing to the right, and then click on Application.
4. Under Storage, find Cookies, and under Cookies, find the option that looks like https://qreports.fas.har...
5. Find the row that's labeled SESSION. In that row, double click the cell under the column Value. Copy it to your clipboard.
6. Create a .env file in your working directory if it does not already exist:

SESSION="..."

7. Paste the value into your .env.
8. When initializing the QGuider, pass in the path to your .env file.

This SESSION key is temporary! You will need to replace it every 30-40 minutes or so if you wish to download any sources. You'll know it's time to replace it when you get 0 QGuide listings when attempting to download QGuides.

Quick Start

import qguider

qgdr = qguider.QGuider(creds=".env")

results = (
    qgdr.query()
    .semesters("Fall 2024", "Spring 2025")
    .download(checkpoint=True, checkpoint_interval=15)
    .parse()
    .agg(by="id") # all courses with the same id will be merged
)

qguider.exporter.to_json(results, "qguider_data/output.json")

API Reference

QGuider

The top-level entry point.

qgdr = qguider.QGuider(creds=".env", outpath="qguider_data")
query = qgdr.query()

• creds — path to a .env file containing credentials
• outpath — directory where downloaded HTML files are stored (default: qguider_data)

Query (fluent builder)

Chain filters before downloading:

Method	Description
.semesters("Fall 2024", ...)	Filter by one or more semesters
.subjects("CS", "MATH", ...)	Filter by subject code
.departments("Computer Science", ...)	Filter by department name
.instructor_last_name("Smith")	Filter by instructor last name
.search("algorithms")	Free-text search
.progress(rich_progress)	Attach a Rich progress bar
.outpath("path/")	Override output directory

After setting filters, call:

# Download HTML files to disk
.download(checkpoint=True, checkpoint_interval=15, report_failed=True)

# Parse previously downloaded files
.parse(skip_failed=True)

# Download and parse in one step
.run(checkpoint=True, skip_failed=True)

QGuideSet

download().parse() returns a QGuideSet, a list-like container of QGuide objects.

len(results)          # number of QGuides
results[0]            # access by index
for guide in results: # iterate
    print(guide.course.title)

# Merge entries that share the same QGuide ID (e.g., multi-instructor courses)
merged = results.agg(by="id")

Data Models

Each QGuide contains:

Field	Type	Description
id	str	Unique QGuide identifier
course	Course	Course metadata
response_rate	ResponseRate	Survey response counts and ratio
course_feedback	CourseFeedback	Likert ratings for overall course, materials, assignments, etc.
instructor_feedback	list[InstructorFeedback]	Per-instructor Likert ratings
hours_per_week	HoursPerWeek	Reported weekly workload distribution
recommendation_strength	RecommendationStrength	How strongly students recommend the course
reasons_for_enrollment	ReasonsForEnrollment	Distribution of enrollment motivations
comments	list[Comment]	Free-text student comments

Course fields: title, subject, department, number, section, instructors, semester, aliases.

Exporting

# Write to JSON file
qguider.exporter.to_json(results, "output.json")

# Return JSON string without writing
json_str = qguider.exporter.to_json(results)

# Convert to pandas DataFrame (requires pandas)
df = qguider.exporter.to_dataframe(results)

Importing

results = qguider.importer.from_json("output.json")

CLI Example

A reference CLI is provided in examples/cli.py:

# Download and parse all semesters, write JSON
python -m examples.cli --download

# Download with Rich progress bar
python -m examples.cli --download --progress

# Parse previously downloaded HTML files
python -m examples.cli --parse

# Import from a previously exported JSON
python -m examples.cli --import

# Skip aggregation of multi-instructor courses
python -m examples.cli --download --no-agg

# Set logging verbosity
python -m examples.cli --download --log-level DEBUG

# Clear all downloaded data
python -m examples.cli --clear-all

Supported Semesters

QGuider currently supports FAS (Faculty of Arts and Sciences) evaluations for:

• Fall 2023
• Spring 2024
• Fall 2024
• Spring 2025
• Fall 2025
• Spring 2026

Notes

• Downloaded HTML files are cached under qguider_data/ by semester, department, and subject — re-running a download with checkpoint=True skips already-downloaded files.
• agg(by="id") merges records that share a QGuide ID, deduplicating instructor feedback and comments across entries for the same course offering.