TracksDebugs

The screen of the app

Inspiration

This was the most down to earth project. Real data for a real problem in an industry requiring more automation without compromising safety.

What it does

Our project extract the information buried in the sensors data to find the anomalies and try to classify them as natural signal variation due to external factor or as possible failure leading to possible unwanted emergency procedures.

How we built it

The core of the application is bases on data mining. Extracting the anomalies using daily data and matching them with previous day data to determine what is noise and what is a possible failure. Then result of the data mining is then displayed in an app for possible preventive maintenance.

Challenges we ran into

The provided data was inconsistent. Lots of data was missing, sometime incorrect. We've never made a map application before.
Splitting the large data to smaller ones to overcome pour limitations of the laptop. Some of us never processed such big and comlex dataset.

Accomplishments that we're proud of

Everyone did the best they could: completed application

What we learned

We improved our capacity to extract meaningful data from noisy and incomplete dataset. Learned to overcome the laptop capacity for mining a large dataset.

The first exposure to the real data (not something prepared for the class etc). Some of us learned how to use PyCharm, pandas, how to clean and pre-process data

MapKit, CSV, work with several UI-elements and usage of big data in the app (iOS developer) Work with different packages and languages in one project (combination of Python and Swift)