Des Moines Public Schools’ (DMPS) Research and Data Management team manages student information and assessment systems and creates reports that serve the entire district.
To create a more accurate dropout prevention model, the team brought five years of student achievement data into Alteryx. The multiple linear regression model—nicknamed the dropout coefficient—weighs student indicators for more accurate reporting. By leveraging this model in Tableau, the Research and Data Management team could extend their analysis with a scatterplot, and quickly identify at-risk students.
Detailed, curated student assessment dashboards allow teachers to dig deeper into this analysis, where they can track student progress across multiple factors.
Today, around 7,000 staff members have access to Tableau Server, so they can monitor performance and ask follow-up questions based on trusted data. With predictive analytics and a central location to access and explore data and dashboards, DMPS better understands how present actions affect long-term student success.
Tableau really made it possible for us to conceptualize where to take this advanced analytics project, because it makes data accessible. Tableau allowed for those conversations around data accuracy, making sure that we're truly looking at the right population for the right reasons.
Insights optimize intervention timelines
Before using Tableau, the district used an Early Indicator System (EIS Report) built in Excel to identify at-risk students. Every six weeks, the team pulled data from the student information system, Infinite Campus, into a spreadsheet. Due to the long reporting process, the data was not timely. Administrators and teachers couldn’t intervene as problems arose.
Originally, the model consisted of 22 indicators, and students had to reach an arbitrary threshold before considered at risk for dropout. And due to the state of the report, the district couldn’t see trends in the data or forecast future roadblocks or opportunities.
“We were waiting for kids to reach the trigger criteria. For example, we were waiting for them to be absent enough to mark ‘yes’ on the absence indicator. As a dropout prevention team, we were being a little bit too late to the game,” said Gene Denny, Database Analyst at Des Moines Public Schools.
The Research and Data Management Team determined that students are likely to show signs of one or more indicators—but some indicators were more indicative of at-risk behavior.
“Our team members felt that it was important that we look at previous years to determine whether or not these indicators were truly affecting the student dropout overall,” said Kimberly Martorano, Data Analyst at Des Moines Public Schools.
To increase accuracy, the team decided to rework the model, weighing the indicators based on data from past years. Additionally, teachers needed more insight into student assessment data to optimize intervention timelines. They needed more capable tools for analysis and communication.
Today, we can have conversations with the administration. With Tableau, we can show them our model and they can understand how factors are affecting our demographic, our population, the type of students that we serve, and we can continually go back to it to extend our analysis.
Linear regression model helps teachers identify at-risk students
The team brought five years of student achievement data into Alteryx to create a multiple linear regression model—nicknamed the dropout coefficient. Instead of 22 indicators, the team reduced the number to the most crucial 12. The model weighs these 12 indicators to determine which students are at risk for dropout.
In Tableau Desktop, the team began by connecting to their model using the Alteryx Web Data Connector. They created a scatterplot that bands at-risk students into groups—intensive, moderate, monitor, and minimal risk—depending on where they fall on a scale of zero to one. The team also encoded the dashboard’s data with color, turning a mark red if a student has greater than a 50 percent chance of dropping out.
If the administration wants to dig deeper, they can drill down into a tabular “student indicator detail.” This displays a comprehensive list of students that meet the criteria for any of the 12 indicators.
At the classroom level, the team created a dashboard with up-to-date data that shows a distribution of students’ standards referenced grades (SRG) alongside their assessment scores. The scatterplot helps teachers identify students with high proficiency, but low SRG scores—allowing them to adjust their instructional methods accordingly.
For additional proactivity, teachers also track daily SRGs in a course grade analysis dashboard. Stacked bar charts show the distribution of SRG scores across each course and topic. This helps teachers quickly identify the students that may need additional intervention or enrichment.
Kimberly said, “Teachers use filters to drill down, look at their population, and identify their differentiated instruction groups. Teachers are able to identify a red group in one area, a blue group in another area, et cetera, they can adjust their instruction immediately.”
The team sets user permissions at the school level using Tableau Server, allowing teachers to gauge scores against other classrooms, while data stays secure. The dashboards connect to the district’s Microsoft SQL Server, so when teachers edit their grade books, the visualizations update in real time.
Customized intervention through advanced analytics
Instead of making assumptions, administrators, interventionists, and teachers now use data to know where to focus their attention.
“Tableau really made it possible for us to conceptualize where to take this advanced analytics project, because it makes data accessible. Tableau allowed for those conversations around data accuracy, making sure that we're truly looking at the right population for the right reasons,” shared Kimberly.
"As a data analyst, I can dive into the student data, whether it's attendance, behavior, or achievement to identify students that are at risk, need intervention, or need additional assistance. This really helps overall student outcomes."
In Tableau, the team elevates the dropout prevention model with visual analytics, making the data relevant, understandable, and actionable. In the future, the team plans to add yearly data to the model, allowing the district to monitor trends over time even more effectively.
“Today, we can have conversations with the administration. With Tableau, we can show them our model and they can understand how factors are affecting our demographic, our population, the type of students that we serve, and we can continually go back to it to extend our analysis,” explains Kimberly.
Assessment dashboards allow teachers to dig deeper into student achievement data in real time. Teachers can drill down into topics within each subject to customize their intervention.
As a data analyst, I can dive into the student data, whether it's attendance, behavior, or achievement to identify students that are at risk, need intervention, or need additional assistance. This really helps overall student outcomes.
“Teachers can dive into individual topics in each class. For example, I am able to analyze a central topic from an English class,” Kimberly shared. “I am able see what percent of students need some help in analyzing text. Next, I can dive even deeper into the data to identify what group of students those are, adjusting my instructional methods accordingly. As an administrator, I can analyze the following questions; Is it a particular teacher? Is it at a particular school? Do I need to adjust my professional development?”
This statistical approach brings added transparency into teaching methodologies and intervention tactics across the district. It has also increased collaboration between staff members. Today, DMPS’ Professional Learning Communities use data to identify effective teaching methods across schools, departments, and courses.