Space Apps Challenge 2017

Background

• The Rafiki team is passionate about local farmers and we have been struggling with the language barrier especially as Africans. So we attempted to face this challenge with our passion and what little programming skills we have

Achievements

• Rafiki aims to be a multilingual AI assistant that recommends articles, maps and datasets based on location and interest/occupation of the user
• Rafiki can communicate through SMS for those that have no access to internet-enabled phones by sending automatically generated statistical data representing the data/map layers requested
• Rafiki utilizes crowd-sourced data from enthusiasts from different parts of world and walks of life
• Because of the language barrier i.e African/Nigerian farmers/users/enthusiasts can't effectively understand scientific terms and datasets produced by top scientists and centers, and thereby rejecting these ideas as a whole, Rafiki intends to Localize all crowd-sourced data and in the long run statistical data, enabling users to effectively take advantage of the data.

Rescources

• African LanguageTechnology Initiative (www.alt-i.org)
• Nasa EarthData https://search.earthdata.nasa.gov/
• Nasa WorldView https://worldview.earthdata.nasa.gov/
• Nasa ECHO API https://api.echo.nasa.gov/echo-rest/
• GIBS https://wiki.earthdata.nasa.gov/display/GIBS/Geogr...
• HAZPOP app https://itunes.apple.com/us/app/hazards-and-popula...
• Google Maps

Challenges

• Poor Internet connectivity
• Implementing Supervised Learning in Python(the crowd-sourcing)
• Interpreting Map data and statistical data
• Translating interpreted data to local languages
• Retrieving data in CSV format using NASA ECHO API, building statistical data visualisations using Seaborn
• Poor power supply
• Limited access to libraries and repositories.