Charlotte School Search: An AI Advisor Built From a Real Parenting Problem

As a mom to a 3-year-old, I recently started researching schools in Charlotte — and quickly found myself jumping between school websites, maps, academic data, and lots of browser tabs.

That sparked a question: 𝗖𝗼𝘂𝗹𝗱 𝗜 𝘂𝘀𝗲 𝗔𝗜 𝘁𝗼 𝗺𝗮𝗸𝗲 𝘁𝗵𝗶𝘀 𝗽𝗿𝗼𝗰𝗲𝘀𝘀 𝘀𝗶𝗺𝗽𝗹𝗲𝗿 for parents?

So I built 𝗖𝗵𝗮𝗿𝗹𝗼𝘁𝘁𝗲 𝗦𝗰𝗵𝗼𝗼𝗹 𝗦𝗲𝗮𝗿𝗰𝗵. 🎓

What started as a simple school finder evolved into three parts:

📍 𝗙𝗶𝗻𝗱 𝗦𝗰𝗵𝗼𝗼𝗹𝘀 — discover nearby schools on an interactive map and filter by level and type.

📊 𝗖𝗼𝗺𝗽𝗮𝗿𝗲 𝗦𝗰𝗵𝗼𝗼𝗹𝘀 — compare schools using official NC performance, growth, math, and reading data.

🤖 𝗔𝗜 𝗦𝗰𝗵𝗼𝗼𝗹 𝗔𝗱𝘃𝗶𝘀𝗼𝗿 — ask questions in plain English and get answers grounded in the same school data.

For now, the app is intentionally restricted to Charlotte-Mecklenburg Schools (CMS) and relies on official CMS and NC DPI data wherever possible.

What makes this project special to me is that it started with a real question in my own life — how do I make a thoughtful school decision for my daughter? — and became an opportunity to explore how AI can make real-world information easier to navigate.

The goal was to build a practical application that helps parents explore Charlotte-area schools through a streamlined interface instead of manually piecing information together.

🔗 Try it here: https://lnkd.in/eFz-23e3

I’d love feedback, especially from Charlotte parents: What else would you want to know when researching a school for your child?

A big shout out to Aishwarya Srinivasan, Arvind Narayanamurthy and The Gen Academy for demystifying so many of the buzzwords around AI.

hashtag#AI hashtag#AgenticAI hashtag#GenAI hashtag#EdTech hashtag#Python hashtag#Streamlit hashtag#CharlotteNC hashtag#BuildInPublic Activate to view larger image,

The core claims in the post hold up well. Charlotte-Mecklenburg Schools is a real, large public district, and the NC DPI performance, growth, and reading/math data referenced are legitimate, publicly accessible datasets — so the app's data foundation is credible. The three-part structure (map-based finder, data comparison, and AI advisor) and the CMS-only scoping are consistent and reasonably framed design choices rather than exaggerated claims. The one thing that can't be verified from the outside is how well the AI Advisor actually performs — whether its answers are reliably grounded in the underlying data — since that requires hands-on testing of the live app rather than just reading the description. The personal narrative (the tab-overload moment that sparked the project) is, by nature, unverifiable but also not the kind of claim that needs fact-checking. Overall, no contradictions or inflated claims turned up, and the content lands at roughly 60% substantive/real and 40% framing, narrative, and promotional elements (hashtags, stylized text, shoutouts) typical of a LinkedIn build-in-public post.