Inspiration
At first we wanted to implement AI to save electricity. But during research, we have realised that the bigger and often miss looked area is human. Linking back to our past internship experiences, we have realised that a lot of things are often overlooked, such as duration of running a server and ways the user can offset them.
What it does
It provides the company with a means of improving their sustainability. When reading through different companies sustainability report, they have often missed out on their staff usage, but mainly targeting their hardware manufacturing usage. Arming with this knowledge, the company can reward/encourage employees to be sustainable.
How we built it
We used Scikit-learn and logistic regression as our model and used Google Colab to train and build the model. For the data, because of a lack of real world dataset, we have to generate our own datasets.
Challenges we ran into
Overfitting of the data. However, we believe that with real-world data, we will resolve this issue. The second challenge is to think of ways to prevent people from cheating.
Accomplishments that we're proud of
We have think out of the box by coming up with a solution to a problem that people may have overlooked. We have also scrap our two other previous ideas and implement this new one in time.
What we learned
The carbon footprints that are generated from training an AI model and offsetting our carbon footprints.
What's next for Sustainable Workplace
We believe this is just the beginning. As you can see, we have used a very simple model and evaluation matrix due to time constraint. But with real-world data, we believe we can observe certain areas that can improve sustainability, while saving money for the company.
Log in or sign up for Devpost to join the conversation.