Exifa.net

Meet Exifa, your AI assistant for understanding EXIF data.

Comment

About Me - Sahir Maharaj

  • As a Data Scientist, I've collaborated with several Fortune 500 companies such as BMW, Volkswagen Group, Audi, SAP, and EY.
  • Creator of AI research application "Wordlit" (400K+ Users in first 3 months) - Featured as #1 app on Streamlit, showcased on newsletters, social platforms, etc.
  • Top 1% Microsoft Power BI Super User
  • Top 1% Data Scientist (Kaggle Master)
  • Creator of AI Newsletter with 30K subscribers
  • Creator of PowerBI.style (Free Power BI Themes with 10K+ downloads)
  • LinkedIn Top Voice Data Science, Machine Learning and AI
  • 200+ Industry recognized certifications (Certified AI Engineer and ML Specialist by IBM)

Inspiration

Exifa is powered by Snowflake

  • Many are unaware of the valuable information stored in image metadata, or EXIF data. I aimed to create a tool that could both educate and assist users in exploring this data.
  • At the same time, I wanted to raise awareness about the potential dangers of sharing EXIF data online.
  • While this data can be incredibly useful, it can also contain sensitive information that you might not want to share publicly.
  • This project is a start of a research exploration to explore EXIF data and find ways to make it both informative and safe for everyone to use.

What it does

  • Exifa analyzes the metadata in your images to provide detailed information about camera settings, colors, and more.
  • It can compare different images, explain technical terms, and even strip metadata for privacy.
  • Users can upload images and ask questions to gain deeper insights.
  • Exifa helps people understand the technical aspects behind their photos.
  • It also has powerful document handling features (RAG).

The goal of the project was to primarily demonstrate Snowflake Arctic and Streamlit technologies.
No third-party APIs have been used.

How I built it

  • I used Snowflake Arctic for data processing, ensuring accurate and efficient analysis of EXIF data.
  • Streamlit was chosen to create an easy-to-use interface that enhances user experience.
  • I focused on developing features such as visuals that allow users to explore and understand their image metadata.
  • I also implemented functionality for comparing images and explaining technical terms to educate users.
  • Privacy features were added to help users strip metadata, addressing the potential risks of sharing EXIF data online.

Challenges I ran into

  • Combining Snowflake Arctic's data processing with Streamlit's interface took patience. I also created an intuitive user experience throughout the app which required a lot of fine-tuning (icons, animations, etc)
  • In addition, making sure the EXIF data analysis was both accurate and fast was tricky, given model hallucinations in a few instances.

Accomplishments that I'm proud of

  • Built Exifa from scratch in just 20 days.
  • Contributed to raising awareness about the privacy risks of sharing EXIF data online.
  • On track to reach a large audience (100K+), as I am scheduling it for promotion in my newsletters, social channels and featuring in my blog.

What I learned

  • Streamlit is incredibly versatile and user-friendly for building web apps (primarily using it in my data science workflows)
  • Snowflake Arctic offers powerful data processing capabilities - Lightweight and excellent
  • Building dynamic features in Streamlit is both rewarding and complex. A real-time user interaction enhances the overall app experience.
  • EXIF data contains a wealth of information that can be easily accessed.

What's next for Exifa.net

  • In talks with Microsoft to publish the app on appsource (more coming soon!)
  • Version 2 to be released in upcoming weeks including integration with third-party APIs, while using Snowflake Arctic as the primary LLM.

Built With

Share this project:

Updates