ClimateJua

• 

  Screenshot of the ClimateJua Application

Inspiration

Living and working in an Arid and Semi-arid area in Kenya called Tharaka, I experience the effects of Climate Change firsthand. I got inspired to tackle the challenge of how we can leverage AI through Partyrock to gain actionable outcomes that can be applied at the community level to mitigate the negative effects of Climate Change.

What it does

ClimateJua works by empowering African Communities in Africa with Byte-sized Climate Knowledge on Climate Change topics by providing bullet point action points that can be expounded further through chat. It also has an Image Generation Feature to make that topic relatable and drive empathy. Jua is a Swahili word meaning "To Know" and when combined with Climate it means to know about the Climate.

How we built it

We built it by generating a prompt to build an application that can help African Households deal with the negative effects of Climate Change in an actionable way

Challenges we ran into

Sometimes the images were not clear or did not accurately depict the Climate Change topic so we had to iterate on the prompt and tweak the settings of the Stable Diffusion Widget. We also realized that the app at first was giving a very small summary of the action item in a paragraph which made it difficult to derive the action items from so we also tweaked the prompt to include bullet points.

Accomplishments that we're proud of

The ClimateJua is providing outcomes that are really relevant and we have configured the app with a chat feature that allows anyone to build on the outcomes and generate action plans and discuss just about anything beyond climate issues.

What we learned

The ability to use the input of one widget to another really made the app function like clockwork. We also needed to learn about how we can generate photographic images and other elements such as seed.

What's next for ClimateJua

With the byte-sized information we are looking into how the knowledge can be localized into vernacular and ensure that we can contextualize with additional data from the different Arid Areas.

Built With

• amazon-web-services
• bedrock
• claude
• partyrock
• stable-diffusion