STANDOG

Inspiration:

Educators are the foundation of society, and having more unified standards produce productive people.

What it does:

Analyzing textbook pages through the provided url's, domains, descriptions, and clusters allowed us to train our KeyBERT model to predict education standards and their associated definitions.

How we built it:

We utilized Python, Google Colab, and KeyBERT. Provided the separation of cells, we were able to cleanly allocate roles toward data cleanup, analysis, visualization, and implementing the model.

Challenges we ran into:

Debugging on a time crunch, not having enough data for better predictions, no correlation between training and testing data- which led us to working with semantic search, and no snacks online made us peckish.

Accomplishments that we're proud of:

Completion of our project, properly implementing a trained model, and fostering camaraderie.

What we learned:

Text vectorization methods like TF-IDF and Bag of Words to learn the steps toward predictive modeling, we learned how to fine-tuning models, KeyBERT, using JSON's, key matching, and furthered our data analytic abilities.