1 of 1

Automating API Management Analytics

​

​

​

Hana Ahmed

University of Michigan ‘27

ITS Data Science Practice

Scan here to contact me or learn more!

Problem

Manually updating key metrics for our API directory was inefficient without centralized data consolidation, making tracking limits and capacity difficult. An efficient, reliable system for accurate, timely updates was needed

​

​

​

​

​

​

​

​

​

​

Methods

Results

  • Accuracy: Provided stakeholders with the most current and accurate data each month, including tracking outliers and anomalies.
  • Enhanced Visualization: The generated graphs effectively visualize progress and growth overtime, aiding in data-driven decision-making.
  • Efficiency: Decreased time spent on the task by 96.7%
  1. Data Retrieval:
  2. Used GitHub REST API, Apigee APIs, Google Sheets API, and TeamDynamix API to extract repository metrics, detailed API data, real-time outage data, and support tickets.
  3. Data Processing:
  4. Developed Python scripts to call APIs
  5. Structured data into Pandas DataFrames for efficient analysis.
  6. Visualization:
  7. Executed Python scripts within Jupyter Notebooks while leveraging Matplotlib to create precise, high-quality visualizations
  8. Automation:
  9. Automated the entire process with GitHub Actions, scheduling monthly updates and committing visualizations to the repository for a consistently updated dashboard

​

INTERNSHIP

Showcase

Goals

  • Automate the retrieval and updating of key metrics to ensure timely and accurate information.
  • Develop a reliable CI/CD pipeline that minimizes manual intervention and reduces errors.
  • Create clear and informative visualizations to enhance data-driven decision-making.

​

This workflow automatically runs on the first day of every month to ensure up-to-date visualization of key metrics related to the API directory. It retrieves the latest data via respective APIs (Application Programming Interface) and generates six visualizations:

  • Total Number of Developers Existing Per Year
  • Total Number of API Calls Per Year
  • Total Number of API Proxies Existing Per Year
  • Total Number of Code Repositories Per Year
  • Total Number of Support Tickets Per Year
  • Total Number of Outages (impacting APIs) Per Year

​

​

I created an automated dashboard that updates and visualizes key metrics for U-M’s API Directory every month. This system ensures that stakeholders always have the most current data without needing to manually update it. The main takeaway is that the automated workflow saves time and effort, providing accurate and up-to-date visualizations that help stakeholders make better, data-driven decisions.

​