Create Your First Project
Start adding your projects to your portfolio. Click on "Manage Projects" to get started
Full Stack Web Application | Temperature and Humidity Reporter
Project Type
Internet of Things, Smart Factory, Web Development
Date
June 2021 - August 2022
IBM Almaden Research Center
This project was developed at the Almaden Research Center for their manufacturing lab.
Lead Engineer
This project was designed and developed by myself, with guidance given by my mentor.
While working for the IBM Tape Department as a software developer intern. Joseph was tasked with solving a pressing issue facing the department's manufacturing line: temperature and humidity monitoring.
Within the manufacturing line, there existed several microclimates that, if left unmonitored, would cause machines to malfunction and occasionally the tape drives that were currently being manufactured to corrode or receive electrostatic discharge damage. Due to the fact that different areas of the manufacturing line operated at different levels of humidity and temperature, a thermostat was not sufficient to guarantee the environment fell within specification. Therefore, I was tasked with developing a solution that successfully monitored these dangerous microclimates to ensure their conditions were desirable to the manufacturing process.
The solution I developed consisted of three major components, which will be discussed in detail below: an IoT embedded system, full stack web application, and cloud functions that predicted future conditions.
First, I developed an IoT embedded system to collection temperature and humidity data at important locations within the manufacturing line. The embedded system consisted of Raspberry Pi 0 W boards interfaced with SHT3X sensors that collected temperature and humidity data every 5 minutes. I wrote several scripts in Python that accomplished this and several other fail safe measures that allowed the boards to continue to collect and store data if something went wrong, such as the internet connection going out. This data was sent to a DB2 database hosted on IBM cloud for storage.
Next came developing a full stack web application that allowed engineers in the department to view the current temperature and humidity data. I developed the backend using Django, and wrote the front end using React. With this full stack web application, users were able to see a live feed of temperature and humidity data, as well as the locations the data was recorded. The data was displayed in several forms, such as line charts and heat maps.
Lastly, came developing a method that would alert users when these microclimates started to exhibit undesirable conditions. Although the web application is useful for live updates and visuals of the current data, it was imperative that there was something put in place that could not only alert users when the environment was out of specification, but could analyze the current data and predict future outcomes. Using SARIMA, I developed an algorithm that would routinely analyze the data collected, and send alerts to operators if dangerous trends were detected, several hours before the environment could fall out of specification. I used IBM Code Engine to host these functions on the cloud, and operators were notified through both email and a Slack workspace message if dangerous trends were ever predicted.