The Journey of Our Data
From raw data to revealing discoveries, follow the transformative path of how we offer real data-driven solutions for any campaign.
1. Raw Data Collection
The secretary of state department of each state releases the raw voter file. This data, depending on the state, is updated roughly every month. Each row represents a voter like you, paving the foundation of our journey.
2. Geo-encoding
Utilizing the provided address details, we take the data to the next level by integrating geographical coordinates, which unlocks the potential for sophisticated geospatial analysis. By leveraging a state-of-the-art method known as geo-encoding, in conjunction with the reliable United States census API, we ensure precise and actionable insights for our users.
3. Join with Geospatial Datasets
By joining our data with public geospatial datasets, we gain a deeper spatial context about the surrounding region of each voter. Different datasets, including urban density, census information, and other geodatabase information allows us to find robust insights about each voter
4. Data Cleanup
We refine the dataset by eliminating anomalies, ensuring accuracy and relevance. This is done by making sure all variables are appropriate for machine learning models
5. Machine Learning
We implement AI and machine learning algorithms in order to extract patterns from the dataset. Specifically, we predict three metrics that determine elections: Voter turnout likelihood, Republican likelihood, and Democrat likelihood. These three variables are scored from 0 to 1.
6. Discovery of Trends & Analyses
With our refined data, we are equipped to uncover meaningful trends and analyses that allow us to predict trends in voting behavior and election outcomes. Our data and methods are open source for anyone.