Track: Social Good

[Disaster Relief] [Best Data Visualization Hack] [Best Social Impact Hack] [Best Hack for Building Community]

Inspiration

Do you ever worry about what is in your water? While some of us live in the luxury of getting clean water for granted, some of us do not. In Flint, MI, another main water pipeline broke. Again. Under water boil advisory, citizens are subject to government inaction and lack of communication. Our goal is to empower communities with what is in their water using data analytics and crowd sourced reports of pollutants in tap water found in people's homes.

What it does

Water Crusader is an app that uses two categories of data to give communities an informed assessment of their water quality: publicly available government data and crowd sourced tap water assessments taken from people's homes. Firstly, it takes available government data of blood lead levels tested in children, records of home age to indicate the age of utilities and materials used in pipes, and income levels as an indicator of maintenance in a home. With the programs we have built, governments can expand our models by giving us more data to use. Secondly, users are supplied with a cost-effective, open source IOT sensor that uploads water quality measurements to the app. This empowers citizens to participate in knowing the water quality of their communities. As a network of sensors is deployed, real-time, crowd-sourced data can greatly improve our risk assessment models. By fusing critical information from these two components, we use regression models to give our users a risk measurement of lead poisoning from their water pipelines. Armed with this knowledge, citizens are empowered with being able to make more informed health decisions and call their governments to act on the problem of lead in drinking water.

How we built it

Hardware: To simulate the sensor system with the available hardware materials at HackMIT, we used a ESP32 and DHT11 Temperature and Humidity Sensor. The ESP32 takes temperature and humidity data read from DHT11. Data is received in Nose-RED json by specifying HTTP request and the sending actual post in the Arduino IDE.

Data Analytics: Using IBM's Watson Studio and AI development tools, data from government sources were cleaned and used to model lead poisoning risk. Blood lead levels tested in children from the Center for Disease Control was used as the feature we wanted to predict. House age and poverty levels taken from the U.S. Census were used to predict blood lead levels tested in children from the Center for Disease Control.

Challenges we ran into

  1. We are limited by the amount of hardware available. We tried our best to create the best simulation of the sensor system as possible.

  2. It was hard to retrieve and clean government data, especially with the need to make data requests.

  3. Not all of us are familiar with Node-RED js and there was a lot to learn!

Accomplishments that we're proud of

1.Learning new software!

  1. Working in a super diverse team!!

What's next for Water Crusader

We will continue to do research for different water sensors that can be added to our system. From our research, we learned that there is a lead sensor in development.

Built With

Share this project:

Updates