💡 Inspiration
We often take convenience for granted. The cool breeze an air conditioner provides, or the joy of flooring the accelerator is intoxicating, but the Earth has to face toxic externalities.
While we have many campaigns and awareness schemes, they can seem too generalized or superficial. Our project aims to achieve the best of both worlds - to show the user exactly what he/she/they owe to our planet and reward them accordingly for doing the suggested actions.
This would help in achieving what we call Targeted Conservation, something that could add to General Conversation to help Guard the Earth
⚙ What it does
We have implemented an array of software and hardware tools to identify and predict energy consumption of a user, including temperature and humidity sensors, and by extracting weather data and using automobile mileage data, we estimate the number of trees/alternative methods to remedy this energy usage.
By harnessing soil data, we can harness the power of machine learning to effectively suggest the ideal plant for the user to plant.
After doing so, they earn TreeCoins, which they can then redeem in stores (easily accessible through our dedicated map) for eco-friendly merch.
🔧How we built it
- For the Machine Learning part, we used open data, performed EDA on it, and trained a classifier on it. We used numpy, pandas, and scikit-learn for the same.
- For the map visualisation, we used leaflet.js through folium.
- For frontend of website, we used HTML, CSS and Vanilla Javascript.
- We used a flask backend that underpins the whole project.
- For the hardware, we used a nodeMCU with DHT11 Temperature and Humidity Sensor
💪 Challenges we ran into
- The project is very big and distinct, lots of merge conflicts and issues were faced to stitch everything together
- Using a hardware and communicating it with ML engine and a webpage was a new challenge for us
📌 Accomplishments that we're proud of
- We were able to completely make the website, hardware, fixed bugs in the given time which seemed impossible
The hardware captures your location which is an added advantage to get results based on your location (check out the Learn more tab in the website)
(This map means the hardware is situated in this location)We are proud to achieve an accuracy of 98.9% in modal over randomized samples
📚 What we learned
- We learned how Machine Learning can be used with hardware inputs
- How to handle multiple tech stacks from different domains and maintaining sync and avoiding any conflicts when everything is summed up
⏭ What's next for Guarden
- Authentication to store user sessions
- Integration of earthcoin into a full-on blockchain
- Increasing Merch library
Built With
- css3
- flask
- html5
- iot
- javascript
- nodemcu
- python
Log in or sign up for Devpost to join the conversation.