Upon encountering the challenge "Everything Starts with Water", our team immediately recognized it as our mission. As our team is made of both: educators and students, we were determined to develop a solution that would benefit both these groups. Our journey began with an extensive exploration, which involved engaging with numerous high school and college students to identify the prevalent issues concerning water-related information. Subsequently, we reached out to educators and seasoned hydric engineering professionals to gain comprehensive insights into the matter. Through these conversations, one core issue emerged: Accessibility.

During our discussions with high school students, a recurring theme emerged: they struggled to access in-depth and complex data related to water, particularly when it concerned topics like Climate Change and global water flowpath.

Educators, on the other hand, emphasized the difficulty of finding comprehensive water-related information in languages other than English and use this data on a way that their students would understand.

Even for professionals in the field of hydric engineering, like Denise Brunoro, that we interviewed for the challenge, visualizing data without a background in data science proved to be a challenge, often necessitating extensive hardware resources and technical know how.

Armed with this knowledge, our team set out to define our primary objective: How can we create a user-friendly platform that not only presents information but also makes data readily accessible and interactive for students and educators worldwide?

After talking with students, It quickly became evident that a particular solution consistently stood out as the most useful and user-friendly for them: ChatGPT, or other Large Language Models (LLMs). This discovery was particularly exciting for us, as it held the promise of overcoming the language barrier by enabling us to deploy the model in multiple languages, since most of these models already supports multiple languages.

However, we recognized that we still needed a way to make data more accessible and visible. To achieve this, we embarked on a comprehensive exploration of the resources at our disposal, including those provided by NASA and other agencies. After diligently assessing various sources such as NASA Earth Data, NASA SWOT, NASA STEM, and NASA GRACE, we set to work on creating the most intuitive interface for students: The Earth Globe.

Our approach involved leveraging a 3D Earth Model, the NASA Worldwind, to collect and integrate the wealth of data we had obtained from NASA and other space agencies into this dynamic and interactive map.

What was once a complex process has now been streamlined into a simplicity itself – just ask "Katara", our Language Model, and let the AI guide you through the 3D Earth Model effortlessly!

With Katara by your side, there's no need to possess the skills of a data scientist to access and visualize up-to-date information. Simply type your question in our chat, request a map display, and Katara will seamlessly transform complex data into engaging and informative visualizations. Needs to check the biggest lake on the world? In a couple seconds Katara will answer you, and show on the map at the same time! It's as simple as that!

The interface we designed and built during the NASA Space Apps Challenge 2023!

WorldWindJs was used for displaying the interactive 3D globe, which is a fork of an open-source planetary globe engine, the Web WorldWind platform. Thanks to the NASA Worldwind documentation, we discovered the optimal solution for presenting and utilizing all the other data sources we had collected. This enhancement significantly improves the experience of interacting with the Earth Globe.

We employ a standard Earth Globe Map from Worldwind as our base map, and we dynamically adjust the map layers in response to user requests. For instance, if a user wishes to view all the world's rivers, we activate a layer containing that data. Similarly, if a user wants to visualize water waste, we render a different layer. Currently, we are utilizing the following layers for Demo:

HydroWASTE Source

HydroWASTE is a comprehensive global database containing 58,502 wastewater treatment plants (WWTPs) along with their associated characteristics.

The database integrates national and regional datasets with additional information to complete missing details, such as the population served, effluent flow rates, and wastewater treatment levels.

Link to HydroWASTE

HydroRivers

HydroRIVERS is a vectorized line network representing all global rivers with a catchment area of at least 10 km² or an average flow of at least 0.1 m³/sec, derived from HydroSHEDS core layers.

It includes 8.5 million individual river reaches with an average length of 4.2 km, covering a total of 35.9 million km of rivers worldwide.

Link to HydroRIVERS

HydroLakes

HydroLAKES offers shoreline polygons for all global lakes with a surface area of at least 10 ha.

It encompasses 1.4 million individual lakes or reservoirs, covering a total surface area of 2.67 million km², a shoreline length of 7.2 million km, and a storage volume of 181,900 km³.

Link to HydroLAKES

We plan to further enhance these maps with EarthData Observer datasets.

Our team has never worked with Geo Processing before, it was our very first time and it was pretty exciting, we also used GeoServer to render these layers in our WorldWind earth model!

Katara serves as the knowledgeable assistant who can answer any water-related questions for both students and educators. Furthermore, Katara acts as a guide through our interactive map, which now offers a comprehensive repository of information about water. This includes data on rivers, lakes, water usage, and even the impact of climate change on water resources.

To built the Katara Language Model we didn't limit ourselves to solely utilizing the resources available through NASA's "Everything Starts with Water" Instead, we diligently gathered data from a multitude of NASA texts and sources. Our objective was to refine our language model to transform Katara into an expert in Water Resources and Climate Change impact. Additionally, we sourced data from various space agencies and organizations.

Here are some of these sources:

- [All NASA "Everything Starts with Water" text resources]

- [Arctic Sea Ice 6th Lowest on Record; Antarctic Sees Record Low Grow...]

- [Arctic Sea Ice Minimum | NASA Global Climate Chang]

- [Carbon Dioxide Concentration]

- [Climate Change Evidence: How Do We Know?]

- [Global Surface Temperature]

- [Ice Sheets | NASA Global Climate Change]

- [Methane | Vital Signs]

- [NASA Climate Kids]

- [NASA Global Climate Change: Sustainability and Government Resources]

- [Ocean Heat Content | NASA Global Climate Change]

- [Overview | Mission - AIRS]

- [Solimões and Negro Rivers]

- [The Causes of Climate Change]

- [The Effects of Climate Change]

Given the diversity of our data sources, we have compiled comprehensive details in our repository.

Also more details about Katara Language Model can be found here.

While we faced time constraints that prevented us from harnessing the full potential of satellite data from sources like NASA STEM, NASA GRACE, SWOT, and Earth Observer datasets, our enthusiasm knows no bounds. We have ambitious plans to gather this data and bring our globe to the forefront of real-time updates.

What's even more exciting is our invitation to professionals from all walks of life to join hands with us. We aspire to transform this project into an open-source educational platform, accessible to every student and educator worldwide. This initiative not only grants access to our language model but also provides a dynamic 3D map system for a richer learning experience. Together, we're forging a brighter future for education and exploration!

Check our teaser video:

https://youtu.be/5msuBrq2Mpg

And our Slides:

https://www.slideshare.net/RaulGuedes6/katara-project-nasa-space-apps-2023pdf

Portuguese Version:

https://youtu.be/wf9irlKhoEo

GitHub Repo

https://github.com/project-katara/

Full Demonstration Video

https://www.youtube.com/watch?v=rtFM8AYH_T0

2023 NASA Space Apps Challenge

Team Katara