Extrapolating climate data and crop calendars
There is a lack of localized agricultural information, especially in the languages spoken within the poorest communities of the world.
INSEAD’s Toto Agriculture initiative addresses this issue. The challenge put forward for the hackathon relates to the problem that climate data is available for a limited number of places around the world. For some (remote) villages, therefore, we need to “estimate” climate by extrapolating “proximate” data. INSEAD will provide Toto Climate Data and Crop Calender Data and online expert guidance.
The problem to be addressed
INSEAD’s Toto Agriculture initiative addresses the issue of a lack of localized agricultural information. INSEAD will provide Toto Climate Data and Crop Calender Data and has formulated two challenges:
Crop calendar challenge (across space and species)
INSEAD has climate data for a number of places around the world. For some (remote) villages, therefore, they need to “estimate” climate by extrapolating “proximate” data.
- if a crop calendar is missing for a specific village, what would it likely be given what we know of similar agro-ecological or proximate zones?
- if a species is missing, what is the likely crop calendar given that we know what other species have as calendars in that location?
Climate challenge (across space)
Objective: if there is no climate polling stations in a village, what is the likely climate of that village across a number of variables (temp, rainfall, dew points, etc.), over time (month by month) that is more accurate than simply using the closest location as a proxy.
Code skills: Kriging or other advanced geo-spatial extrapolation methods, C# or .net.
A zip file will be provided containing 2 versions of INSEAD’s TotoGEO 2.0 Climatic data. Format #1 is an Access 97 .mdb file containing a very flat table – It will have the location, and the climatic information for that location – where each row contains a different type of climatic data. For example, one row may contain temperature maximum for a location, and the next may contain sunshine days.
The second data set is a MySQL 5.6 (5.6+ required) table backup of 3 tables that are normalized. They define shapes to the nearest climatic information site. INSEAD used voronoi polygons in QGIS to pre-calculate the nearest sample, creating shapes that can be indexed easily by MySQL. It makes querying the climate information at any location very quick. An example query on this data would look like:
select * from totoclimate_shapes a inner join totoclimate_metrics b inner join totoclimate_fields c on c.field_id=b.field_id and b.metric_id=a.metric_id where ST_CONTAINS(a.shape, POINT(@longitude, @latitude)) order by c.`sort`;
Also crop calendar data will be provided, in MS ACCESS format (raw data, to be extrapolated from).
Inspiration and relevant links
Koppen logic and/or FAO definitions of zones
A large survey was conducted by WorldAgInfo (http://worldaginfo.org/) which noted the lack of localized agricultural information, especially in the languages spoken within the poorest communities of the world. A gap between information producers, and those needing that information the most was identified. While some countries are blessed with well funded Ministries of Agriculture, most do not have the resources to fill the “content gap”. Largely supported by the Bill and Melinda Gates Foundation and INSEAD North America, a series of projects involving content automation platforms have been launched to increase access to agriculture knowledge to the world’s poorest areas. Organizations involved in these projects include Plant Resources for Tropical Africa (PROTA), CABI, ISRIC – World Soil Information, Africa Soil Information Service, the Grameen Foundation, Farmer Voice Radio, Farm Radio International, the GSM Association, and Technoserve, among others.
The initiative is part of a larger project that is distributing (via SMS, MMS, video, and radio) agricultural tips, health tips, weather reports, and related information to rural areas, especially to persons at the very bottom of the income pyramid. Starting with a small group of partners in Malawi, Kenya, Tanzania, and Uganda, the project is growing to a number of other countries. Additional components can be found at www.totoagriculture.org.
Contact challenger (name, email, Skype, mobile)
Crop Calender Data:
Zaw Moe Sann