| Hello, *|NOME|* In March, IMDS organized a workshop at the Municipal Department of Education of Rio de Janeiro to present the tool “School Dropout Antecedent,” also known as Predictor. The meeting worked as a dynamic listening and roundtable conversation, creating a dialog between the two institutions about the necessary mechanisms to get the Project running already during the 2025 school year. The idea was to bring together people who work in different areas of the Department of Education to explain how the tool works and add the vision of those teams to the work of the institute. The School Dropout Record is a tool capable of anticipating which students – enrolled in Junior High School (6th to 9th grade) in the municipality's public network – ¬¬ are at risk of not achieving the learning objectives and, therefore, of becoming very prone to abandoning classes. The tool was presented at the end of last year to the Municipal Department of Education of Rio de Janeiro and aroused the interest of some of the subnational entities with which IMDS has a Technical Cooperation Agreement (TCA). Dialoguing with the actors of the education network is an essential step towards ensuring that the tool meets the predominant needs of schools. These are the professionals who experience the challenges of the classroom and school management on a daily basis. Their direct involvement also allows important adjustments in the usability and functionalities of the project, increasing the chances of the tool being useful day after day, as well as that it be adhered to. The workshop began with an explanation of the science that underlies the project, based on the use of advanced machine-learning techniques, and the exposure of the data used in the tool by IMDS data scientist Monica Bahia. She points out that the driving force comes from the concept of “studying the past to try to predict the future.” The model is based on the available histories of students in the school system and seeks to answer a central question: With all the clues that the student himself leaves, is it possible to anticipate whether he will face obstacles to advance in the course? Clues left in previous years help to get this up-to-date forecast. The Dropout Antecedent analyzes the available data about a student and the school he attends. Furthermore, it is well known that school management is important and influences the students’ results. The Predictor uses what is meant by the “first-day model” – a survey and anticipation of important information in the first days of the school year. Throughout the year, the device is fed with more data as the student takes tests and their engagement in the classroom is observed. The results obtained in the first and second two months are taken into account, stages that work as a thermometer to measure the performance and results of the efforts to alleviate obstacles. In a second moment, researcher Bruna Goussain conducted a practical activity with the professionals who participated in the workshop, based on the interaction with the Predictor Usage Map. The following questions guided the dynamics: What is done today to avoid failure? When and how would you use the tool? How would you like to view this data? How can the device help with decision-making? What are the possible challenges to adopting the tool? The conversation focused on the possible actions and obstacles faced by managers due to the debate that arose from the initial questions. During the dynamics, the discussions addressed the importance of knowing the risk factors that affect students and how the tool can act in this analysis. In addition, the group highlighted the principal of a school as being the key player in this process. The professionals argued that if the director understands public policy and how it works, in addition to understanding what is possible to do with the data presented, he can move the pieces of this structure, putting into practice what is developed in the secretariat. Another relevant dialogue was about the difference in the socioeconomic profile between schools in Rio, which implies a significant difference in performance. This variation is noticeable between regions, and the distance does not need to be great for differences to appear. Two schools in the same neighborhood can have different profiles, depending on the location – whether they are inside or outside a community, for example. The elaboration of the Dropout Antecedent is one of the fruits of the partnership between IMDS and the Municipal Department of Education of Rio, and the implementation of this tool emerges as an innovative solution in the fight for better quality learning. Good learning is a key element towards increasing the chances of social mobility for the most vulnerable students in our country. IMDS is fully focused on developing tools that can improve the performance of these students. See you in the next "IMDS Letter"! Paulo Tafner CEO | 
