AI and Data Science Academic Programs in the United States
Description
This dataset was created to organize and analyze information about Artificial Intelligence (AI) and Data Science (DS) academic programs offered by universities in the United States. The goal was to build a structured database that shows not only general university information, but also how programs, admissions, and outcomes are connected. The dataset includes five main tables: University, UniversityYearStats, DegreeProgram, AdmissionRequirement, and ProgramOutcomeStats. The University table contains basic details such as the school name, location, whether it is public or private, and student population. The UniversityYearStats table was created to track changes over a five-year period (2021–2025), including tuition, rankings, and program growth. The DegreeProgram table focuses on specific AI and Data Science programs, including degree level, focus area, duration, and tuition. AdmissionRequirement includes common requirements such as GPA, GRE, and TOEFL, while ProgramOutcomeStats provides general outcome information like enrollment, graduation rate, employment rate, and average salary. To build this dataset, I used publicly available information from the U.S. Department of Education College Scorecard and official university websites. Since there is no single dataset that contains all program-level details, some information was gathered manually and organized into a consistent format. All data was cleaned and structured in Excel before being imported into MySQL Workbench. The dataset was designed using relational database principles and normalized to Third Normal Form (3NF), which helps reduce redundancy and keep the data organized. Each table represents a specific type of information and is connected through primary and foreign keys. This makes it easier to run queries, analyze relationships, and expand the dataset in the future. Overall, this dataset can be used for learning database design, practicing SQL queries, or exploring trends in AI and Data Science education across U.S. universities.
Files
Steps to reproduce
To create this dataset, I started by gathering real university data from the U.S. Department of Education College Scorecard website. I downloaded the full dataset and opened it in Microsoft Excel. From there, I filtered the data to focus only on relevant columns such as university name, city, state, control type (public or private), student population, and tuition. I selected around twenty universities, making sure to include schools from different states and a mix of both public and private institutions. After that, I worked on collecting program-level information. Since there isn’t one dataset that contains all AI and Data Science programs, I used official university websites to look up programs such as Data Science, Artificial Intelligence, and Computer Science. I recorded details like program name, degree level (Bachelor’s, Master’s, or PhD), approximate tuition, and duration. I then organized this information in Excel and made sure each program was linked to the correct university using a unique ID. Next, I created admission requirements based on common criteria I saw across multiple universities. Most programs listed things like minimum GPA, GRE requirements, and English proficiency tests such as TOEFL. I standardized these into a consistent format so they could be stored in a separate table. To include changes over time, I built a table for university statistics over five years (2021–2025). Since not all data was available year-by-year in one place, I used general trends from tuition and enrollment data to create a consistent structure. I also added fields for number of AI and Data Science programs to show how offerings may grow over time. For program outcomes, I included fields like enrollment, graduation rate, employment rate, and average salary. These were based on typical values reported by universities and were organized so they could be analyzed across programs. Once all the data was collected, I cleaned everything in Excel by removing extra columns, renaming headers, and making sure data types were consistent. I also created unique IDs for each table to support relationships between them. I then moved to MySQL Workbench to design the database. I created the tables, defined primary keys and foreign keys, and made sure the structure followed Third Normal Form (3NF) to avoid redundancy. After that, I imported the CSV files into MySQL and tested the database using SQL queries to make sure everything was connected properly. Finally, I used the reverse engineering tool in MySQL Workbench to generate the E-R diagram and confirm that all relationships were correctly set up. Tools used in this process included Microsoft Excel for data cleaning and organization, MySQL Workbench for database design and implementation, and publicly available sources such as College Scorecard and university websites for data collection.