Staralytics

Inspiration

The project was born out of the realization that Stara's current fleet fuel consumption analysis is entirely spreadsheet-driven and has a lot of room for improvement and machine-computed insight. When speaking with the team and mentors from Stara, we quickly received excited feedback about the idea of improving the analysis of Stara's spending, fuel consumption, and CO2 emissions and connecting their data with predictive analysis. After this we worked quickly to analyze the most exciting and valuable insights that could be automated and extracted from Stara's open data.

What it does

As it stands today, the system

extracts maintenance patterns
notifies Stara when vehicles are approaching periodic maintenance checkups
breaks down fuel consumption data
summarizes fleet CO2 emissions

About the team

We're a group of students at KTH, studying Software Engineering of Distributed Systems. Check out our projects at our website.

How we built it

Using a Flask backend backed by PostgreSQL and various Python modules, we built a server-side API that ingests Stara's various data streams and cross-correlates them in useful ways.

The front end is driven by Angular and provides a responsive and easy-to-use dashboard with graphs, charts, and tables bringing out new insights into how the company can allocate spending and optimize their fleet of city vehicles.

Challenges we ran into

The most painful challenge was the presence of many outliers in the data streams provided by Stara.

Accomplishments that we're proud of

In a short amount of time we were able to pull together a vast amount of data into previously unexplored metrics that can help inform Stara about future optimizations. The team worked well across the dashboard, separating tasks elegantly and in a very organized fashion.

What we learned

There are always trade-offs to be made when designing real-world applications aimed at providing the most business value. During the design and development processes, we had to account for unexpected setbacks and difficulties, but were able to pull through with what we deemed to be the most valuable and interesting data analyses.