Inspiration
The inspiration for Kiliwatch stemmed from the growing concern for sustainable development in urban environments especially in Kilimani area in Kenya. With the increasing emphasis on reducing carbon footprints and integrating green building practices, it became essential to leverage technology in making informed decisions. The project aims to provide developers, contractors, and planners with crucial insights into their construction projects' environmental impacts, ensuring strategic and effective planning.
What it does
Kiliwatch tool evaluates construction projects' environmental impacts by analyzing development plans either through manual entry or by parsing uploaded documents. It assigns environmental scores to buildings based on criteria such as age, size, insulation, energy sources, and green features, while integrating external data like land surface temperature, wind speed, rainfall, and canopy height. The tool provides interactive visualizations, AI-generated insights, and actionable recommendations for improving sustainability. It offers real-time performance and comprehensive reporting, aiding urban planners, developers, and contractors in making eco-friendly decisions and promoting sustainable building practices.
How we built it
Geospatial Data Acquisition
QGIS for Building Footprints : We used QGIS, an open-source geographic information system, to extract detailed building footprints from satellite imagery. This helped in accurately determining the spatial extent of each building within a project area. ArcGIS for Wind Speed Data : ArcGIS's extensive library of climate data provided critical wind speed information, which was integrated into the environmental impact analysis to assess potential wind-related effects on building performance and energy efficiency. Google Earth Engine for LST and Canopy Height : Google Earth Engine allowed us to perform high-resolution analysis on Land Surface Temperature (LST) and Canopy Height. By leveraging its powerful computational capabilities, we were able to analyze and incorporate temperature variations and vegetation data into the project assessments. CHIRPS for Rainfall Data : CHIRPS (Climate Hazards Group InfraRed Precipitation with Station data) provided reliable rainfall data, which was critical for evaluating water management features and assessing the overall sustainability of development projects.
Frontend
Streamlit : The entire frontend is built using Streamlit, a powerful Python library that enables the creation of beautiful, interactive web applications with minimal code. Streamlit simplifies the integration of data visualizations, forms, and interactive components. Plotly : For rendering advanced data visualizations, we utilized Plotly's Python library. This allowed us to create interactive charts, statistical plots, and map visualizations efficiently. Folium : Folium was used to create interactive maps, helping users visualize buildings, environmental scores, and spatial data intuitively.
Backend
Data Processing and Computations : Rasterio : For reading and manipulating geospatial raster data such as LST, Rasterio provided the necessary tools to handle large datasets and perform spatial analytics. Pandas and Numpy : These libraries were essential for data manipulation, analysis, and performing statistical computations. Azure OpenAI :We integrated Azure OpenAI to generate detailed analysis and recommendations. The AI model processes the project data and provides actionable insights based on predefined prompt templates.
Challenges we ran into
Ensuring the accuracy and consistency of geospatial data from diverse sources like QGIS, ArcGIS, Google Earth Engine, and CHIRPS required rigorous validation and preprocessing steps. Developing robust methods to handle multiple file formats (PDF, DOCX, XLSX) for extracting relevant project details was challenging due to varying document structures. Rendering large datasets and interactive visualizations efficiently while maintaining responsiveness was a significant concern. Continuous optimization was necessary to enhance user experience. Designing an intuitive and user-friendly interface that caters to both technical and non-technical users required iterative improvements based on feedback.
Accomplishments that we're proud of
We're proud of successfully using Kiro to aid in creating the application and integrating our Google Earth Engine (GEE) model with the application to develop a highly useful tool.
What we learned
Working with Land Surface Temperature (LST) data enhanced our understanding of handling geospatial information and interpreting remote sensing data. We realized from the statistics obtained by the model, Azure OpenAI could be able to generate insights and recommendations which would allow the non-technical people to understand the project. Parsing various formats of project files (PDF, DOCX, XLSX) and extracting meaningful information showcased the multifaceted aspects of data extraction and management. We got to use Plotly and Streamlit which significantly impacted the way we presented the data-driven insights .
What's next for KILIWATCH
Next steps for KILIWATCH include enabling the creation of detailed development plans based on previous ones in the area, enhancing the model for analysis beyond Kilimani to support various geographic regions, integrating more comprehensive datasets for improved accuracy, refining the user interface for better user experience, leveraging AI and machine learning advancements for more precise assessments and recommendations, and seeking user feedback to ensure the tool meets evolving needs. These enhancements aim to make KILIWATCH a versatile and powerful tool for sustainable urban development globally.
Log in or sign up for Devpost to join the conversation.