Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/holzerjm/civichacks-demo/llms.txt

Use this file to discover all available pages before exploring further.

The EduHack dataset (eduhack_boston_schools.txt) is a synthetic educational equity report for Boston Public Schools, 2024-2025. It contains detailed metrics on enrollment demographics, MCAS achievement gaps, chronic absenteeism, technology access, teacher staffing, and college readiness.

​
Dataset overview

File: data/eduhack_boston_schools.txt
Scope: Boston Public Schools Educational Equity Report, 2024-2025
Size: ~51 lines, 22,000+ characters
Format: Plain text, structured report

​
Data sections

​
District overview

  • 49,152 students across 125 schools
  • 78 home languages represented
Demographics:
  • Hispanic/Latino: 43.8%
  • Black: 29.1%
  • White: 14.2%
  • Asian: 8.9%
  • Multiracial: 3.2%
Special populations:
  • Low-income families: 62%
  • English Language Learners: 32.4%
  • Special education services: 21.7%
  • Experiencing homelessness: 4.8% (2,359 students)

​
Sample queries

Here are the three built-in queries for the EduHack track: 
python scripts/demo_step2_rag.py edu 1

​
Query 1: What are the most significant achievement gaps in Boston public schools?

This query surfaces:
  • Math: White-Black gap of 38 percentage points (52% vs. 14%)
  • ELA: White-Black gap of 41 percentage points (63% vs. 22%)
  • ELL students: 8% math proficiency, 5% ELA proficiency
  • Low-income vs. non-low-income: 34-point math gap, 34-point ELA gap
  • No improvement in 5 years for White-Black math gap

​
Query 2: How does transportation affect student attendance and outcomes?

This query identifies:
  • Students with 45+ minute commutes: 2.3x more likely to be chronically absent
  • District-wide chronic absenteeism: 38.2%
  • High schools: 47.3% chronic absenteeism rate
  • Strongest predictor of absenteeism in the data
  • Superintendent recommendation to expand transportation options

​
Query 3: What technology access barriers exist for students and teachers?

This query highlights:
  • 18% unreliable home internet, 7% no internet at all
  • 3,400 eligible families still unconnected despite Comcast partnership
  • 34% of 6th graders lack digital literacy to navigate educational software
  • 45% of teachers feel unprepared to integrate AI tools
  • 1:1 device ratio achieved but connectivity gaps remain

​
Key metrics reference

Student groupProficiencyGap vs. WhiteGap vs. Overall
Asian58%+6 pts+30 pts
White52%Baseline+24 pts
Overall28%—Baseline
Low-income17%-35 pts-11 pts
Hispanic/Latino16%-36 pts-12 pts
Black14%-38 pts-14 pts
ELL8%-44 pts-20 pts
Students w/ disabilities6%-46 pts-22 pts
GroupRateNotes
Students experiencing homelessness62%Highest impact
High schools (avg)47.3%5 schools >55%
Black students44%—
Hispanic/Latino students41%—
District-wide38.2%Up from 33% pre-pandemic
White students28%—
Asian students22%Lowest rate
45+ min commute2.3xStrongest predictor
MetricValueContext
Device ratio1:1One Chromebook per student
Unreliable home internet18%—
No home internet7%—
Comcast Essentials enrolled8,200 familiesPartnership program
Eligible but unconnected3,400 familiesGap remains
6th graders lacking digital literacy34%Cannot navigate software
Teachers unprepared for AI tools45%Professional development gap
MetricValueContext
Total vacancies (SY2024-25)1275.2% vacancy rate
Special Education vacancies42Highest need area
Math vacancies28—
Science vacancies22—
Bilingual Education vacancies19—
Students of color64%Black + Hispanic/Latino
Teachers of color39%Demographic mismatch
Roxbury turnover28%Highest in district
Mattapan turnover25%Second highest

​
Using this dataset in the web app

When you select EduHack in the Gradio app (Step 3), the interface displays:
  • Header: ”📚 EduHack — Analyze Boston public schools equity, achievement gaps, and student outcomes data”
  • Example questions: All three queries above as clickable examples
  • Chat responses: AI answers grounded in the specific metrics from this dataset
Try asking follow-up questions like “Which schools have the highest absenteeism?” or “What interventions does the superintendent recommend?” to explore the data interactively.

​
Querying from code

Here’s how the RAG pipeline loads and queries this dataset: 
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settings
from llama_index.llms.ollama import Ollama
from llama_index.embeddings.huggingface import HuggingFaceEmbedding

# Configure local AI
Settings.llm = Ollama(model="llama3.1", request_timeout=120.0)
Settings.embed_model = HuggingFaceEmbedding(model_name="all-MiniLM-L6-v2")

# Load the EduHack dataset
data_file = "data/eduhack_boston_schools.txt"
documents = SimpleDirectoryReader(input_files=[data_file]).load_data()

# Build vector index
index = VectorStoreIndex.from_documents(documents)
query_engine = index.as_query_engine(streaming=True, similarity_top_k=3)

# Query
response = query_engine.query(
    "What are the most significant achievement gaps in Boston public schools?"
)
response.print_response_stream()
See Step 2: RAG with civic data for the full implementation.

Build docs developers (and LLMs) love

Get started for freeTalk to us