Elowane
commented
6 years ago For example DataForGood France made a prediction algorithm for the Red Cross to optimize the food distribution in 800 centres in France. The code and explanation is on their GitHub account here
The explanation is in French unfortunately! To summarize, the Red Cross do not distribute prepared meals but canned food, milk, vegetables etc that they get from supermarkets or private persons for example, and in some centres they may have more than necessary while in others they do not have enough, and this may vary along the year. So the initial scope of DataForGood was:
- to analyse the existing mesh of distribution centres and determine the zones with a food deficit and the zones with a food excess (using statistical data about family revenue).
- to analyse the local partnerships with supermarkets and determine where new partnerships would be beneficial
I'm not sure yet if they managed to do that or changed the scope along the way, but anyway what was interesting for French Red Cross may be different than for Norwegian organizations.
For the organizations like Røde Kors, Fattighuset or Frelsesarmeen who house and feed homeless or poor people, it may be interesting to be able to predict how many people will come so that they can be better prepared, depending for example on the date or on the weather. The first step would then be to contact them to check if they have past data, and if they would be interested in a prediction tool!