CScope | Devpost

Inspiration

What it does

How we built it

Elevator pitch: CScope empowers users to fight climate change by providing them with personalized data and climate-based solutions. Users can calculate their personal carbon footprint, get climate related recommendations from a GPT-3 based chatbot, and access data on increasing global temperatures through an interactive temperature map.

Inspiration

Climate change is one of the most significant challenges of our time, and the need for action has never been greater. CScope was inspired by the urgent need to empower individuals to take action against climate change. By providing access to critical data, tools, and resources, CScope aims to help people understand the impact of their actions and make informed decisions that reduce their carbon footprint. Our goal is to make it easy and accessible for everyone to contribute to the fight against climate change, so we can create a more sustainable future for generations to come. We chose the name CScope to suggest that our product provides a broad view of climate patterns, much like a telescope provides a view of the cosmos, encouraging users to explore and investigate the facts on their own in addition to incorporating natural language technology to suggest further ways to get involved.

What it does

[MAP] Using UC Berkeley’s Climate Change: Earth Surface Temperature Data, we created an interactive temperature map that users can use to find temperatures of countries around the world. The map also has the ability to predict temperatures into the future as well .using a seasonal automotive regression model.

[CLIMATE COMPANION] ClimateCompanion is a personalized chatbot which is tailored towards climate change and the purpose of this feature is to allow users to ask questions they have about improving their daily habits or developing new habits to reduce their carbon footprint and help the environment. We have a calculator which gives the users their current carbon footprint and this Chat bot will act as the user’s companion in guiding them towards a sustainable life. For instance, if users see that their car is harming the environment a lot, they can ask “How should I use my car in a way that reduces carbon emissions?” to the Climate companion and it will give the users tips regarding the same.

[CARBON FOOTPRINT CALC] The Carbon Footprint Calculator helps users track their carbon emissions. The calculator takes in information such as users' household usage, flight trips, car miles driven, as well as buses and trains taken to create a comprehensive carbon footprint value for that category. Users can even enter specific information such as kilowatt hours of natural gas usage, natural gas, and coal to get a more accurate breakdown of their carbon footprint depending on their usage. When giving input, users have a choice for the units they want to enter the information in, this allows our application to be more accessible. The calculator also gives the sum of all carbon emissions calculated in the different categories on a tab called “Results” which allows the user to see their total carbon footprint.

How we built it

[HOW THE MAP WORKS] The map is built using geopandas and a dataset from UC Berkeley. It is integrated into our application along with a year slider to provide temperature information over time.

In order to predict future temperatures, we used the Python library statsmodels’s Seasonal Autoregressive Integrated Moving Average, or SARIMAX, model with parameters (3,0,0) to represent the lagging (past values), differencing (this is what makes non-stationary data stationary), and white noise (for modeling shock events) trained on 75% of the total data (observations up till the mid 1800s). Each country’s model had a root mean squared error of no greater than 3 degrees, the full details of which can be found in the country_scores.txt file. This was later integrated into the map via [INTEGRATION WITH MAP]

[CLIMATE COMPANION] ClimateCompanion was created using OpenAI’s GPT-3 API, specifically the text-davinci-3 model. The chatbot is tailored specifically using OpenAI’s api to respond specifically to climate change related questions. We made it tailored to our needs by adding informational prompts to the user input.

[CARBON FOOTPRINT CALC] The carbon footprint calculator was made using Carbon Footprint Ltd. API. We integrated their calculator service into our app which allows our users to evaluate their current carbon footprint in order to ask the Climate companion appropriate questions to reduce it.

Challenges we ran into

…

One challenge we ran into was integrating all the different API’s into one cohesive application. In addition, identifying the appropriate model to train the temperature predictions was a challenge. None of the members of our team have trained time series data before, a type of data that often features serial/autocorrelation and thus cannot be accurately modeled with ordinary least squares regression. Thus, several of the sources cited below were an enormous help in demonstrating how to accurately model such data with Autoregressive Integrated Moving Average model.

Accomplishments that we're proud of

We are very proud of being able to integrate all of these developed features into one functional application, which helps users find everything in one place to make their lives more sustainable.

Additionally, we’re proud of being able to make a working model for each country to predict future temperatures within a couple degrees! Each model took a significant amount of time to train and tune and such information can be incredibly useful across the world.

Overall with CScope, we've accomplished our goal of providing individuals with the tools and resources they need to fight climate change. Our powerful temperature map, carbon footprint calculator, and GPT-3-based bot have enabled users to access critical data and make informed decisions to reduce their impact on the environment. By empowering individuals to take action against climate change, we're helping to create a more sustainable future for all. With CScope, we're proud to have made a significant contribution to the fight against climate change.

What we learned

Through the development and implementation of CScope, we've learned several important lessons about the fight against climate change. First, we've learned that access to data and resources is critical for individuals looking to reduce their carbon footprint and contribute to the fight against climate change. Second, we've learned that personalized recommendations and actionable steps are essential for empowering individuals to take meaningful action. Third, we've learned that collaboration and integration with other climate tools and resources can enhance the effectiveness of individual efforts. Finally, we've learned that ongoing updates and improvements are necessary to keep pace with evolving climate trends and the needs of users. Overall, our experience with CScope has taught us that addressing climate change requires a multifaceted approach that engages individuals, organizations, and governments alike.

What's next for CScope

There are several improvements that could be made to CScope to enhance its functionality and user experience. One possible improvement would be to expand the temperature map to include more recent data, allowing users to track the impact of climate change in real-time. Additionally, the carbon footprint calculator could be updated to include more detailed data on specific activities, making it easier for users to understand their impact on the environment. For example, the calculator allows users to add kilowatt hours of electricity. However, it would be beneficial to relate this to more relatable metrics such as hours charging a laptop or number of dryer cycles run. Another possible improvement would be to enhance the capabilities of the GPT-3-based bot, allowing it to provide even more personalized and actionable recommendations for users looking to reduce their carbon footprint by using account information to keep track of user-specific information (such as geographic location, dwelling type, dietary restrictions, etc). Finally, CScope could benefit from greater integration with other climate tools and resources, providing users with a more comprehensive and streamlined experience.