Beyond Weather Forecasts: How We Built JalTaap with AWS to Turn Environmental Data into Action
At the WeMakeDevs × AWS Bharat Builds Tour hackathon at DTU, Team DevCore built JalTaap, a multilingual environmental safety platform that combines location-specific risk assessment, voice accessibility, and automated issue reporting. Here’s a look at the architecture and engineering decisions behind the project.
# Beyond Weather Forecasts: How We Built JalTaap with AWS to Turn Environmental Data into Action
Weather data is everywhere. Weather applications can tell us the temperature, show rainfall predictions, and display alerts. But a more important question remains: **How do we turn that information into decisions that help people stay safe?**
That question shaped our work on JalTaap, a project developed by Team DevCore during Environmental Hacks 2026, part of the WeMakeDevs × AWS Bharat Builds Tour at Delhi Technological University (DTU).
Our objective was to build more than another weather dashboard. We wanted to create a platform that could interpret environmental conditions, communicate the risks in a user's preferred language, and help communities report problems to the authorities responsible for resolving them.
This article explores the technical thinking behind JalTaap, the AWS technologies supporting it, and the lessons we learned while designing a solution around a real-world problem.
## 1. Starting with the Right Problem
One of the first challenges was deciding what information users actually needed.
A temperature reading by itself does not tell a construction worker whether conditions are suitable for strenuous work. Similarly, a general rainfall warning may not reveal whether a particular route is affected by waterlogging.
We identified three gaps:
- **Context:** General weather information does not always translate into guidance for a specific location and activity.
- **Accessibility:** Information may be difficult to use when it is available only through text-heavy interfaces or unfamiliar languages.
- **Actionability:** Reporting a flood, leak, or water shortage does not guarantee that the report reaches the responsible department.
These gaps became the foundation of our design. Instead of treating weather, accessibility, and civic reporting as separate problems, we brought them together in one platform.
## 2. Designing the JalTaap Workflow
JalTaap follows a straightforward process: collect a request, assess the relevant conditions, communicate the result, and support action.
**Step 1: Collect the user's request**
Users can interact through WhatsApp, Telegram, or the website. They can type a question, share their location, or send a voice note. This reduces the need to navigate complex menus before receiving useful information.
**Step 2: Assess environmental conditions**
The platform retrieves forecast information and evaluates relevant factors, including temperature, humidity, hot nights, rainfall, air pollution, and ultraviolet exposure.
An important engineering decision is that a tested calculator determines the risk level. AI helps with conversation and image assessment, but the core heat-risk calculation does not depend on a language model guessing the answer.
**Step 3: Deliver understandable guidance**
JalTaap converts the result into practical information, such as safer working hours, hydration guidance, and nearby cooling facilities. Its conversational experience supports Hindi, Hinglish, and English, with voice-based interaction for users who prefer speaking and listening.
**Step 4: Connect reports to action**
Users can report issues such as flooding, water leaks, and tanker shortages. Reports can be checked, mapped, and routed according to location and problem type, while the relevant authorities receive a structured view of the issues requiring attention.
This workflow is the central idea behind JalTaap: environmental information should lead to an understandable next step.
## 3. The AWS Architecture Behind the Project
Building this workflow required a combination of compute, storage, AI, communication, scheduling, and security components.
We used AWS services and open-source tools to organise these responsibilities.
### AI and conversational capabilities
Amazon Bedrock supports AI-powered conversations and image checks. Strands Agents supports the agent workflow, while Amazon Transcribe and Amazon Polly support voice-input and spoken-response capabilities.
The goal is to make the system easier to interact with without allowing AI to replace the deterministic calculations used for heat-risk assessment.
### Serverless processing
AWS Lambda provides serverless execution for backend operations, while Amazon API Gateway supports API access. Amazon SQS supports asynchronous processing, and Amazon EventBridge Scheduler supports scheduled tasks and alerts.
This approach helps separate different responsibilities and provides a foundation for scaling workloads without managing traditional servers.
### Data and file storage
Amazon DynamoDB supports application data, while Amazon S3 provides object storage for relevant files and reports.
These services help the application manage information generated through user interactions and reporting workflows.
### Web experience and security
AWS Amplify supports the web experience, and Amazon Cognito supports authentication. AWS CloudWatch provides monitoring capabilities.
We also used AWS Serverless Application Model (AWS SAM) for infrastructure-as-code workflows, AWS Lambda Powertools for operational practices, and Cedar for access-control policies.
Each component has a specific purpose in the architecture. The value comes from how these components work together to support the complete user journey.
## 4. A Key Engineering Insight: Heat Is More Than Temperature
One of the most significant technical considerations in JalTaap is humidity.
High humidity reduces the body's ability to cool itself efficiently. This means potentially dangerous conditions can occur even when temperatures remain below a conventional 40°C heatwave threshold.
We tested our humid-heat rules against modelled Open-Meteo weather data from April to July 2026.
Our analysis identified **414 dangerous humid-heat city-days across 63 Indian cities that a simple 40°C threshold would have missed.**
In Bhubaneswar, our analysis identified 50 dangerous days, of which the temperature-only threshold caught three. In Kolkata, it identified 42 dangerous days, none of which met that threshold.
These results demonstrate why risk assessment needs to consider more than a single temperature value. They are based on modelled weather data rather than direct weather-station measurements, so they should be interpreted accordingly.
For our team, this analysis helped connect the technical design to the problem we wanted to address.
## 5. Designing for Accessibility
Accessibility was not simply a visual-design decision. It influenced how users communicate with the system and how information is delivered.
JalTaap uses short responses, visual indicators, and spoken guidance. Its multilingual conversational experience is designed for people who may not be comfortable navigating a conventional application or reading long explanations.
The ability to use Hindi, Hinglish, or English makes the interaction more flexible. Voice notes allow users to communicate naturally, while spoken responses reduce reliance on reading.
The broader lesson is that an effective interface should adapt to its users rather than expecting every user to adapt to the interface.
## 6. Connecting Personal Safety with Civic Reporting
JalTaap also addresses a different but related challenge: what happens when users encounter environmental problems in their communities?
A report about a flooded street, leaking water supply, or missed tanker delivery needs more than a place on a map. It needs a path towards the organisation responsible for taking action.
Our reporting workflow is designed to check incoming reports, associate them with locations, and route them to relevant departments. An authority-facing view helps organise reports into queues and action lists.
This approach creates a link between individual observations and administrative workflows.
Although routing and verification can improve how reports are handled, real-world resolution still depends on data quality, departmental processes, and the response of the responsible authority.
## 7. Testing Beyond the Demo
A successful demonstration is only one part of building a dependable system.
Our project documentation records 196 automated tests passing and 34 test conversations used to check AI responses. We also tested the risk calculations against published reference values.
These checks helped us evaluate different parts of the application, from deterministic risk logic to conversational behaviour.
However, automated testing cannot guarantee that every forecast, report, or recommendation will be correct under every real-world condition. Weather values may differ from observed conditions, and local flood information may require human verification.
We designed JalTaap as a supplementary safety tool, not a replacement for official weather warnings or emergency services. This distinction matters when developing technology that can influence real-world decisions.
## 8. Lessons from Building JalTaap
Working on JalTaap reinforced several engineering principles.
**First, start with the user's decision, not the technology.** The purpose of the platform is not to display more data; it is to help users understand what to do with that data.
**Second, assign the right responsibility to each component.** Deterministic calculations handle risk assessment, while AI supports communication and image-based workflows. This separation helps make the system easier to reason about and test.
**Third, treat accessibility as part of the architecture.** Language selection, voice interaction, and simple visual design affect the entire user journey, not just the appearance of the website.
**Finally, design for the full lifecycle of information.** A useful system must collect information, process it, communicate it clearly, and provide a path towards action.
## 9. What We Want to Improve Next
Our next priorities include integrating official meteorological warnings when the required access is available, expanding language support to more Indian languages, and exploring SMS access for users who cannot rely on WhatsApp.
We also see opportunities to improve the quality of local flood information and develop partnerships with employers and city control rooms.
These improvements would help us move from a working technical solution towards broader real-world adoption.
## Conclusion: Engineering for Impact
JalTaap began with a simple observation: access to weather information does not automatically mean access to safety.
By combining location-specific environmental assessment, multilingual communication, voice accessibility, and civic reporting, Team DevCore developed a platform intended to turn data into practical guidance.
The WeMakeDevs × AWS Bharat Builds Tour hackathon at DTU gave us an opportunity to explore how cloud services and thoughtful engineering can address a problem that affects people's everyday lives.
For us, the most important takeaway is this: **good engineering is not just about building a system that works. It is about building a system that people can understand, trust, and use.**
*JalTaap was developed by Team DevCore for Environmental Hacks 2026, Heat & Water track, as part of the WeMakeDevs × AWS Bharat Builds Tour.*
*Technical note: Heat-risk analysis in this article uses modelled Open-Meteo weather data for April–July 2026. JalTaap is a supplementary safety platform and does not replace official meteorological warnings or emergency services.*
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article