Link to the datastory: click to this link
Note for a good visualization: for a correct user-experience, we recommend to use a zoom of 67% on a 13 inches (Tested on a 13 inches MacBook Pro).
Notebook containing the project is project.ipynb. As jupyter notebook doesn't reproduce the embedded HTML, the already rendered version is available HERE
Despite recent technological improvements, there is still a large share of people who are starving. “1 in 9 people in this world go to bed on an empty stomach” (source: World Food Program). The aim of this project is to propose a way to end starvation in Africa, one of the continents most affected by this hunger issue. To do so, we will tackle the problem from two different perspectives. On one hand, the analysis will show an estimatation of which African countries have food deficits. On the other hand, the investigation will give the reader a complete picture of the food availability across Europe. The FAO dataset will be intensively used as it contains useful information about net food supply for both Africa and Europe. The analysis section will then leave space to a more important chapter of our story: how to solve the uneven distribution of food? We would like to reallocate resources that otherwise would be end up as food waste from the richest European countries to the poorest African ones. This combined effort approach is a simple solution we think could potentially have a huge impact on many lives!
-
How can we quantify hunger in Africa in terms of kcal needed per capita per day? What are the regions that need the most help?
-
Can we make truthful predictions on African and European population for 2020 based on data we have until 2013? And what about food supply? Will our predictions be accurate?
-
By analysing the time period 1960 to 2020, could it be possibile to assess if the African food situation actually improved and why? In this context, can we explain important changes in food availability observed by correlating them with historical facts?
-
How Europe evolved from 1960 to 2020 in terms of food waste?
-
How can we quantify food surplus in Europe? What are the regions characterised by the highest food waste rates?
In milestone 3:
-
Can we produce a list of simple and highly nutrient-dense final products that can be shipped from Europe to Africa?
-
How can we define a consistent model to predict which European countries have to ship and how much should they individually contribute?
-
Which countries can provide the resources to meet this plan (based on previously observed data)?
-
If everyone does work together, can we solve hunger on this continent?
Food balance
This section contains food supplies for every country in the world. As the documentation reports, a region supply is defined as: “Production + imports - exports + changes in stocks (decrease or increase)”.
We will use this database to analyze the amount of kcal/person/day for each African and European state in the food categories we are interested in.
Prices
As a primary resource, we used the FAO dataset to obtain prices of the food items analysed in our diet. In this context, FAO did not contain all the prices needed.
GDP
GDP was retrieved and used in order to obtain an overview on how much every European country should contribute to the cause.
Population
The dataset was retrieved at this link. This dataset was used to determine the male and female population for Europe and Africa from 1950 to 2020 with a granularity of 5 years. We will interpolate to obtain data with 1 year frequency.
All FAO’s dataset are in .csv so easy to parse and to work on. Same applies to United Nations dataset. The dimensions are overall manageable and no problems emerged.
This dataset was retrieved from Kaggle at the following link. The dataset contains the geometry of every country in the world. The dataset can be imported easily as it's a json file and also quite small. The Json file will be then converted to a GeoPandas Dataframe with geopandas (specialized library to work with geographical visualization).
This dataset linked was used to get domestic food prices for items whose prices were not included in the FAO dataset.
(additional dataset found online)
We will use this dataset to build the optimal diet. This dataset contains nutritional information for raw products, covering all those present in the FAO dataset. The information reported are per 100 g of servings.
Estimated calories needed: we will scrape this page to obtain information on the average calories needed for the African population
-
Elaborate a model by which we can calculate an ideal daily caloric intake for men and women. Goal would be to scrape the Estimated calories webpage in order to obtain substantial data on estimated energy needs that we can then compute and adapt to our particular case.
-
Use data from POPULATION section (FAO Database) to extract current African population based on gender.
-
Compute for every region in Africa an estimation of kilocalories demand taking into account the composition of population.
-
Focus on FOOD BALANCE section (FAO Database) in order to retrieve data about African food supply (kcal/ per capita per day).
-
Given the data we gathered so far, we can now compute the amount of kilocalories we have to hit in order to fill the gap. For the sake of simplicity, we suggest to consider a share of 55% of carbohydrates, 25% of proteins and 20% of fat. In this context, we will also elaborate an analysis on the hunger issue on a country basis.
-
We focus now on figuring out a feasible solution for the problem. The first step would be to come up with a draft diet based on simple products (chicken, meat, rice, etc..) that can fulfill the demand in each African country.
-
The USDA ARS food dataset will be used to convert kilocalories (proteins, carbs and fat) into kg of final product we would have to deliver.
-
The above mentioned diet will be the basis for a more specific analysis on the nations (restricted to Europe, USA and China) more suitable to provide the resources. In particular, we are going to rank countries based on price/kg and internal supply of every product.
For the last week we will focus on visualizing the data in the best way and we are thinking about creating:
- An interactive geographical visualization of the poorest country in Africa in terms of lack of food.
Methodology: A proposal will be made how food could be redistributed from Europe to Africa, not countrywise. Even though a country might overall be in a surplus, it could potentially lack some macronutrient (making up for it with more of another one). As we are not only interested in supplying the necessary calories but also providing a healhty diet, the analysis will be continued with into the three macronutrients (Carbs, Proteins and Fat). Due to calories not being a tangible unit for most readers (especially when using orders of magnitude of Gigacalories), representative products for each macronutrient will be introduced to facilitate conveyance for a broader audience. Also, this representation will help us to choose "representative product" to be used in our diet. In a second step the European countries merged as a whole will be reconsidered as unique nations, and the amount of supply they should provide will be decided.
-
Split the nutrition supply into macronutrients (55% Charbohydrates, 25% proteins, 20% fat, source and analyze respective surplus and deficit. The strategy we will adopt is similar as the one we have shown in the first two parts of our notebook. The division into macronutrients is needed as in reality we would have to actually feed people healthily and respecting natural demand of everyone's body.
-
After obtaining such informations, we are able to choose which items are considered as "of interest" to represent the abstract notion of calories. The choice of this final products will be weighted on many factors among nutrient-density, availability in different European countries, etc. As above mentioned, we are committed to create a simple diet composed by few products that combined together will be able to meet African food demand while respecting the contraints we have from European availability of spare food.
-
We now move on aggregating European and African countries and, by representing calories/macronutrients as food items, we decide how much of which food items Europe, considered as a whole, should provide based on what Africa actually needs.
-
After deciding these quantities, we'll find which countries should provide and receive which food items and in what quantities, based on their surplus and deficits, respectively. The goal here is to have a simple but effective way to communicate to the reader order of magnitudes on play in this problem. We will provide real examples on how just one country or a group of countries could sustain African calories' debt (if their population is comparable). In addition, it will be considered to compute in a more accurate way how european countries will have to contribute to fill up the demand. We still have to figure out an efficient model but the idea is to define a share of final products based on individual country's population.
In the last two weeks we will focus entirely on the creation of an effective datastory.
-
Come up with visualization methods for these informations, in order to catch the attention of the user. We have described carefully in the notebook different type of interactive plots we have in mind.
-
Propose to the reader the results of our model on two levels. The first one is aimed to catch the reader attention by showing him just few but interesting results coming from our first analysis on Africa and Europe over the years. In this part, we will use heatmaps, heatmaps with time, short animation of different kinds. As the data story goes on, they reader will gain more knowledge on African history of famine and European food wasting trend. As the second level comes in, our proposal will be to introduce the model of macronutrients, final products and finally reallocation of resources we proposed.
Manuel: Neural networks to predict data of different datasets up to 2020. Analysis on the African population and deficit. Analysis on food prices and product to find the best ones. Analysis on both Europe and Africa. Website maps, chord plot, sunburst plot, layout introduction. Continuosly update on the readme and final reviews on the work.
Joao: All Europe analysis and comparison with Africa. Code refactoring for milestone 2, with function rewriting and implementation. Website design, layout, counter, caloric needs for male and female and difference, scatter plot for GDP analysis.
Dario: Animation on caloric decifit for the notebook. Analysis on Africa supply. Writing all the textual description for the datastory. Preparing the final presentation and poster.
Riccardo: Analysis on Africa supply and comparison with Europe. Coding the maps for the notebook. GDP analysis for the best 5 countries. Coding up the optimization part to obtain values on distribution of food and diet, 3D scatter plot of macronutrients. Continuosly update on the readme and final reviews on the work.