AI Early Warning System Predicts At-Risk K-12 Students

Critical Issue

A large school district in the Western US was facing challenges such as high student dropout rates and a significant number of students not graduating on time. The district struggles with identifying which students need intervention due to limited resources and complex socioeconomic factors, including family income, race, and gender, that impact educational attainment. As a result, there is a pressing need for a targeted approach to allocate support and resources effectively to improve academic outcomes and address the issues contributing to student dropout rates within the district.

Customer Profile

Large school district in Western US

74,000 K-12 students at 90+ schools

Key Problems

High student drop out rates

Too many students not graduating on time

our solution

An early warning system was developed to proactively identify students at risk of academic challenges, starting as early as kindergarten, in order for guidance counselors to intervene and prevent these issues from escalating. The system continually learns and incorporates new risk factors over time. A user-friendly dashboard highlights which students are most at risk and provides intuitive explanations for the risk scores. The intervention team was trained on how to effectively utilize the system. The implementation process took 14 weeks with a team of 4 individuals, and the system is now fully operational. The technology stack used for this project includes Azure, Python, Azure Data Factory, ADLS, AutoML, and Databricks.

The results

Higher graduation rate

Through early identification of students at risk of academic challenges, the district was able to reach their goal of a 95% graduation rate.

improved student insight

The implementation has resulted in a better understanding of students by teachers, guidance counselors, and administrators.