Google Challenge - Team Nordlichter

Inspiration

Road crashes in the European Union claimed in 2019 about 22700 lives and left more than 1.2 million people injured. We want to decrease this number!

What it does

Our solution provides municipalities information about hazardous events so that they can better understand where and why accidents happen throughout their city in order to prevent them in the future. In a first prototype we created a dashboard which explains the correlation between potholes and hazardous events.

How we built it

We fetched telematics data from moving vehicles in the city of Rome and combined it together with open street map data and results from a machine learning model which is able to detect potholes. With this information we built an aggregated dataset and on top we created a dashboard in PowerBI, making the data exploratory.

Challenges we ran into

We faced the challenge of joining our data since locations were stored in different formats. It was possible for us to face this challenge by transforming the location to geohashes, making the data mergeable.

Accomplishments that we're proud of

• Building a presentable prototype in a short time period
• Seeing our hypothesis confirmed that potholes correlate with hazardous events ## What we learned
• Combining telematics data and open street map data can lead to very interesting insights!
• Collected some experience in building dashboards in PowerBI
• Learned how to include data from various sources to build a huge data set ## What's next for Google Challenge - Team Nordlichter Integrate further data into our model in order to identify more patterns that cause hazardous events throughout the city. We also want to include amenities in these areas like schools, crowdy areas with a lot of cafes, restaurants, etc. We also had the idea to look for patterns which are possibly related to remaing trip distance, trip duration or weather data.

Built With

• powerbi
• python