The Kanda Weather Group is seeking participants to develop a simple User Interface (UI) dashboard that shows our forecasts in an easy-to-understand way for local farmers and stakeholders.

The company is refining a weather balloon (also known as radiosonde) IoT technology product that collects data and uses Machine Learning to make a simple 12-hour rain forecast. They are 80% cheaper than traditional radiosondes and can be set up anywhere on earth.

We see Africa as being one of the biggest potential users of our technology, and we are looking to work with local groups to make localized forecasts for small-holder farmers. An accurate 24-hour flood forecast and warning system may be enough time to allow them to gather their crop yields on 1 hectare of land early before dire flooding occurs. Additionally, we believe that the data from the radiosonde balloons can also be used to make short-term air quality forecasts over polluted regions like the Niger Delta or fast-growing metro cities.

We are working with decentralized climate company dClimate to provide near real-time access to other variables such as soil moisture or rainfall data for our two forecast regions Accra, Ghana and Uyo, Nigeria. We invite participants to be creative in their efforts to build a display dashboard, and are awarding $1,000 USD to the 1st place project team and $500 USD to the 2nd place finishers.

About Kanda Weather Group

Our company sees an opportunity in the miniaturization of computing and low cost of IoT sensors to dramatically improve weather data collection of the upper atmosphere. We originated as an idea in the Telos community through the Worker Proposal System and have been able to develop our product independently thanks to Goodblock.io . We are working with universities in Nigeria, Ghana, and the U.S. to collaborate on research.

We are passionate about eosio blockchain, and see low power, long range open networks as holding the potential to bringing many people out of poverty with low cost access to information.

Submission Directions

On July 1st, the Kanda Weather Group will host a live virtual weather radiosonde launch to kickoff the hackathon.

Also on that day, a series of "Hindcasts" will be provided that participants can use as input to their forecasting dashboard. For the purposes of the hackathon, the only difference between a Hindcast and a forecast is essentially the date. It is essential that a dashboard include at least one of these Hindcasts into the display.

You will receive an email from the hackathon when the Hindcasts are posted to Taikai. Additionally, you can view them in the Updates tab, when they are made available.

Deliverables

The submission must include the following by July 31, 2021 at 23:59 GMT :

1. In English, please provide either ONE of the following describing the technologies you used and what makes your dashboard unique:

1. Link to a 2 minute video accessible via YouTube or other hosting platform in the Project's Description section
2. OR   (if your English is not that good)   greater than 250 word summary in the Project's Description section

2. At least 2 different screenshots of the dashboard you created 

3. Link to code on github with README.md file for how to build and run the software. If the dashboard can handle multiple locations and forecasts, please include how to adjust the backend parameters to achieve this functionality.

Example Approaches

1. The "Simple" dashboard

     -   Handles only one location/date/forecast.

     -   For example:   Uyo, 5/20/21, No rain

2. The Adaptable dashboard

     -   Handles many locations, dates, and forecasts

     -   Forecast adjustable via some backend (csv input file, raw input)

     -   For example:    Uyo, 5/20/21, No rain     or       Accra, 5/22/21, Heavy rain

3. The Exceptional dashboard

     -   Handles many locations, dates, and forecasts

     -   Forecast adjustable via some backend (csv input file, raw input)

     -   Reads in soil moisture for the given location/date from dClimate's API 

         and shows flood risk based on that information 

         (HINT: high 10cm soil moisture values or water runoff values indicate higher flood risk)

     -   For example:    

         Uyo, 5/20/21, No rain, Low flood risk     or Accra, 5/22/21, Heavy rain, High flood risk

Figure 1. Example UI weather dashboards.

Judging Criteria

Forecast is clearly illustrated and dashboard designed nicely (40%)

Ability to handle more than one location and/or date/time  (25%)

Ability to handle wide variety of forecast types (rain, heavy rain, flood risk, air quality) (25%)

Interactivity (for example selecting different language like French and/or Yoruba, dark/light themed)  (10%)

Getting Started

Of course, the first steps are to create a profile here on Taikai and find other participants that are interested in your approach to developing this product. Once you've done that, create a project under this competition and go through the necessary steps.

We recommend using one (or more) of the following javascript frameworks when developing a desktop application.

• Vue.js
• Angular
• React (Create react app)

Examples of weather dashboards using Vue.js in particular can be found here.

Please review the resources section for guides on how to use the dClimate REST API.