Mapping Tanzania

Inspiration

Interesting and meaningful challenge that leverages big data and computer vision technology.

What it does

Mapping Tanzania uses a 30 mile x 30 mile satellite image (6 billion pixels) to accurately map roads and distances not found through common services like Google and Bing maps. This leaves many people with inaccessible access to resources and makes it hard to document and find a disaster in real-time. Our app solves this problem.

How I built it

We used a massive satellite image dataset to construct a large node graph that we leveraged to calculate distance. The entire process took many hours and countless data cleaning and processing iterations.

Challenges I ran into

We realized the limitations many libraries have, causing us to switch our tools multiple times throughout the competition/

Accomplishments that we're proud of

• We managed to build a product that if scaled appropriately could actually be used in a developing country. I found it interesting how we were able to find roads that Google and Bing maps have not yet indexed on their road services.
• Successfully combining 64 8192 by 8192 images into a single image.
• We derived an accurate model to convert GPS coordinates to pixel coordinates.

What I learned

I learned a lot about data science and computer vision technology.

What's next for Mapping Tanzania

We need to work on scaling and extra features.

Built With

• big-data
• bingmaps
• cv
• express.js
• maps
• ml
• node.js
• opencv
• python