anchOUR

Before your community Hits Rock Bottom. By analyzing your club's Instagram account, anchOUR detects engagement decline, explains why it’s happening and gives data-backed ways to recover

Comment

Inspiration

Community & friendship are becoming harder to sustain because interest slowly drifts without anyone noticing. Clubs and community organizations often rely heavily on Instagram to engage members. But they usually realize something is wrong only after likes, comments, and attendance have already dropped. We wanted to build a tool that catches that drift early and helps communities course-correct before they hit rock bottom.

What it does

anchOUR is a community engagement tool for clubs and communities. It scrapes and analyzes Instagram data, such as posts, follower trends and engagement metrics to detect when participation is declining and explain why it’s happening. The platform visualizes engagement decay over time and provides data-backed insights and recommendations to help leaders re-engage members and stabilize their community before it fades.

How we built it

  • Google AI Studio (Gemini) for designing a website

  • Python for data processing and modelling

  • Apify to collect engagement and social data

  • Snowflake as our central data warehouse

  • Streamlit for building interactive dashboards

  • Snowflake Cortex AI + Gemini for insight generation, cluster labelling

  • snowflake-arctic-embed-m to create vector embeddings of text data

  • scikit-learn MiniBatchKMeans for clustering social media post types

Challenges we ran into

This was our first time using Snowflake and Streamlit, which came with a steep learning curve. Connecting data pipelines, deploying Streamlit apps inside Snowflake, and effectively using Cortex AI required extensive experimentation and debugging. We also faced challenges translating our Google AI Studio designs into a real, production-ready implementation within Snowflake.

Accomplishments that we're proud of

  • Building a full end-to-end analytics pipeline from data collection to insights

  • Successfully integrating ML and AI inside Snowflake

  • Creating a clear, visually consistent dashboard that matches our neo-brutalist design vision

  • Turning abstract “community health” into measurable, explainable signals

What we learned

  • How to effectively use Snowflake as more than just a database

  • Applying ML + AI directly within Snowflake, including Cortex AI

  • Creating and deploying Streamlit apps, including Streamlit inside Snowflake

  • Creating bright and eye-catching website using Google AI Studio (Gemini)

What's next for anchOUR

Our next step is full integration with the Meta API to access real-time Instagram engagement data. We were limited during the hackathon because we didn’t have a business or creator account for testing, but this integration will unlock deeper analytics, stronger predictive signals, and more accurate recommendations for community leaders.

Built With

Share this project:

Updates