Minnesota Fabric/Power BI Community

Agenda for this community meeting:
- Introductions
- Presentation - From Bronze to Predictions: A Modern Machine Learning Journey
- More about the presentation:
Chad will be utilizing Microsoft Fabric and PowerBI to pull information from the Congress.gov API to create various machine learning models including the following:

1. Clustering – Grouping bills together on natural topical clusters based on their text content. Uses concept of tokenization.
2. Network Analysis – Creates a sponsorship graph across all active legislators computing centrality measures and running community detection to identify natural collaboration clusters. This surfaces who the key “bridge” legislators are - people whose co-sponsorship connects otherwise separate groups.
3. Prediction – Uses classification models to predict things like a bill’s likelihood of advancing (i.e passing committee, reaching a floor vote) based on features like sponsor, committee, bill type, and cosponsor count.

The current congress(119th) has over 17,000 bills and 165K cosponsor relationships. The data has been pulled into fabric via python notebooks and a medallion architecture has been created to help with these models that is updated nightly. Attendees can expect to gain the following during this discussion:

1. Practical examples of clustering, network analysis, and predictive modeling.
2. Insights into building machine learning solutions with Microsoft Fabric.
3. Techniques for transforming machine learning outputs into actionable insights through data visualization and analytics.

- General Discussion - Any pain points anyone is needing support with currently?

***We need presenters and presentation topics for future meetings. Please reach out with ideas or to volunteer!