Best Countries for Exporting Chickens: Market Study

Data Analysis for the International Launch of La Poule qui Chante.

La Poule qui Chante, a French company specializing in the food industry, aims to expand internationally. As part of this expansion, the management team decided to leverage data analysis to identify potential markets and guide strategic decisions.

The primary objective of this project is to identify target country groups for exporting La Poule qui Chante’s products. To do so, an in-depth analysis of data from the FAO (Food and Agriculture Organization) will be conducted.

Data used from : https://www.fao.org/3/X9892F/x9892f03.htm

Summary

• Data Preparation

• PCA with Component Selection

• Component Interpretation

• Correlation Circles and Projection

• Hierarchical Clustering

• K-means Clustering

• Target Countries

1. Data Preparation

We have three files at our disposal: Population, Food Balances, and Annual Time Series. We begin with feature engineering on these dataframes before merging them to obtain a single file with the variables of interest.

Population

• Objective: Determine the most favorable countries for international expansion based on population size.

• Data Exploration: Selection of countries and their population for the year 2021.

• Data Processing:

  • Removal of the "China" variable to avoid redundancy.

  • Filtering of countries with populations under 5 million to focus on sufficiently large markets.

• Conclusions:

  • 124 countries were selected as potential targets.

  • The average population of these target countries is approximately 62.8 million inhabitants.

  • The population sizes vary significantly, from a minimum of 5.1 million to a maximum of 1.4 billion.

    
    

China and India are two outliers (due to population size) that will be kept in mind for further analysis.

Food Balances (2021)

• Objective: Explore and clean the food balance data of various countries for 2021, focusing on meat production and food availability.

• Data Exploration:

• Data Processing:

  • Creation of a pivot table to display meat production by country.

• Results:

  • Selection and addition of the poultry meat production column in thousand tons.

• 

  Conclusions:

  • The food balance data is now ready for deeper analysis.

  • Extreme values in total meat production, notably in the U.S. and Brazil, might require special attention.

Annual Time Series

• Objective: Explore food security data for 2021, focusing on indicators such as the number of undernourished people, GDP per capita, and political stability index.

• Food Security Data:

• Conclusions:

  • India stands out as a significant outlier with 23,390,000 undernourished people.

  • Countries like Luxembourg, Singapore, and Ireland stand out with high GDP per capita.

  • No outliers were found in the political stability variable.

We have merged our dataframes, and below you can visualize the retained variables (with the non-numeric "Zone" variable added). The four outlier countries (China, India, U.S., Brazil) are included in a separate dataframe.

2. PCA with Component Selection

1. Data Scaling:

   • Using the StandardScaler method to normalize the data.

   • Transforming the data to have a mean of 0 and a standard deviation of 1.

2. Objective of Scaling:

   • Ensure all variables are on the same scale.

   • Allow for fair comparison between variables.

3. PCA Initialization:

   • PCA is a dimensionality reduction method that helps explore and visualize data. The dimensionality of our data is reduced to a specified number of components—here, 10 for our 10 variables.

4. Training the PCA:

   • Training the PCA on the scaled data.

5. Objective of PCA:

   • Identify the directions (principal components) that maximize data variance.

   • Reduce the number of dimensions while preserving as much information as possible.

To determine the optimal number of principal components to choose, we need to examine the variance explained by each component and choose a number that captures sufficient variance while reducing dimensionality. Here, we visualize the variances with a Scree plot to help us choose the appropriate number of principal components.

In blue, we have the variance of each new component, and in red, the cumulative variance.

We see that nearly 80% of the variance is captured by the first three components, and nearly 90% by the first four.

3. Component Interpretation

Component F1: Well-being and Development IndicatorsThis dimension can be interpreted as a global measure of a country's well-being and development. High scores indicate a high per capita food and protein availability, along with political stability and high GDP per capita.

Component F2: Population and Food Self-SufficiencyThis dimension reflects the capacity of countries to be self-sufficient in food and produce their own food. High scores indicate a strong internal food availability. This component is strongly correlated with population size.

Component F3: Economic Dynamism and Trade OpennessThis component is characterized by positive correlations with indicators such as exports, imports, and political stability. These correlations suggest that F3 represents the political and economic influence of a country on international trade. Countries with high exports, solid political stability, and higher GDP per capita are likely to have higher F3 values, highlighting their importance in the global economy.

4. Correlation Circles and Projection

• Component F1: Well-being and Development IndicatorsThe points in the upper right of F1 represent countries with the highest well-being and development indicators. These countries are characterized by high GDP per capita, solid political stability, and abundant food availability. On the other hand, the points in the lower left of F1 represent countries with the lowest well-being and development indicators, likely indicating less developed economies, political instability, and food insecurity.

• Component F2: Population and Food Self-SufficiencyThe points in the upper left of F2 indicate countries with large populations but low food self-sufficiency. These countries may heavily rely on food imports despite their large populations. Conversely, the points in the lower right of F2 represent countries with small populations and high food self-sufficiency, producing enough food for their consumption despite their relatively small population.

5. Hierarchical Ascending Classification

We can now begin our hierarchical ascending classification! We visualized a dendrogram with 10 clusters and selected 6 clusters containing a sufficient number of countries.

Parallel Coordinates Diagram

Let’s examine the second cluster in orange (#1):

• Population (score of 4): This cluster shows the highest average population among all clusters, suggesting it includes highly populated countries.

• Dispo_Kcal mean, Dispo_Prot mean, GDP, Political Stability, Import, Export (scores from 0 to -0.5): These variables show relatively low average values in this cluster compared to others. This may indicate lower food and protein availability, as well as lower GDP and political stability. Additionally, these countries seem to have average to low import and export levels.

• Food Availability, Meat Production, Poultry Production (scores of 3): These variables show relatively high average values in this cluster compared to others, suggesting that these countries have medium to high levels of food availability and meat/poultry production. This may indicate a developed agricultural industry or relatively easy access to food resources.

  
  

We also used a heatmap to visualize this information, as well as box plots for each variable across the 6 clusters.

6. K-means Clustering

The score informs us of the data clustering density within each cluster. The higher the score, the more tightly grouped the data is within the cluster.

We observe that the curve rises quickly from 1 to around 6 clusters, then decreases. Beyond 5 or 6 clusters, each additional cluster does not significantly improve the score. Thus, the ideal number of clusters is between 3 and 6.

Result Analysis

We display the data points and centroids.

• Component F1: Well-being and Development IndicatorsThe points in the upper right on F1 represent countries with the highest well-being and development indicators. These countries may be those with high GDP per capita, strong political stability, and abundant food availability.

  The points in the lower left on F1 represent countries with the lowest well-being and development indicators, likely indicating less developed economies, political instability, and food insecurity.

  
  

• Component F2: Population and Food Self-SufficiencyThe points in the upper left on F2 indicate countries with large populations but low food self-sufficiency. These countries may heavily rely on food imports despite a large population.

  The points in the lower right on F2 represent countries with small populations and high food self-sufficiency. These countries may have a relatively small population compared to their capacity to produce enough food to meet their needs.

Zoom on centroids

Summary table with countries by cluster

Cluster	Description	Country in cluster	More information
0 – blue	• Medium to high well-being and development • Relatively high food availability • Potential demand for food products such as poultry meat • Good political stability	'Russian Federation', 'Indonesia', 'Japan', 'Mexico'	We are removing Russia from the list as the FAO data on the political stability index dates back to 2021. When analyzing outliers in poultry meat production, we find Russia, Mexico, Indonesia, and Japan. Therefore, we can assume strong competition.
1 – orange	• For a smaller population, these countries generally have high well-being and development indicators, with high food and protein availability and an average GDP per capita higher than that of other clusters (except for cluster 3). • Political stability is generally high.	‘Saudi Arabia’, ‘Austria’, ‘Belarus’, ‘Belgium’, ‘Bolivia (Plurinational State of)’, ‘Bulgaria’, ‘Chile’, ‘China - Hong Kong SAR’, ‘China, Taiwan Province’, ‘Costa Rica’, ‘Denmark’, ‘United Arab Emirates’, ‘Finland’, ‘Greece’, ‘Hungary’, ‘Israel’, ‘Kazakhstan’, ‘Malaysia’, ‘Norway’, ‘New Zealand’, ‘Portugal’, ‘Republic of Korea’, ‘Romania’, ‘Serbia’, ‘Slovakia’, ‘Sweden’, ‘Switzerland’, ‘Czech Republic’.	Food availability and demand for quality food products can be significant.
2 – green	- Diversity in economic development and political stability - Market competition varies by country	South Africa, Colombia, Egypt, Iran (Islamic Republic of), Pakistan, Peru, Philippines, Thailand, Turkey, Ukraine, Vietnam	There may be a growing demand for poultry meat in some of these countries due to changing dietary habits and economic growth.
3 – red	- Mature markets with high living standards and strong demand for quality food products. - Intense competition, but they offer growth opportunities. - Good political stability and a high GDP per capita suggest that these are favorable countries for exporting our chickens.	Germany, Argentina, Australia, Canada, Spain, France, Italy, Netherlands, (Kingdom of) Poland, United Kingdom of Great Britain, and Northern Ireland	The high GDP per capita suggests that the market is favorable for deploying our sales strategies in these countries.
4 – purple	- Lower development indicators and variable political stability. - Demand for poultry meat exists but is limited by economic and political factors. - Expansion into these countries would require a thorough assessment of risks and opportunities.	Algeria, Angola, Azerbaijan, Benin, Cambodia, Congo, Ivory Coast, El Salvador, Ecuador, Ghana, Guatemala, Guinea, Haiti, Honduras, Jordan, Kyrgyzstan, Lebanon, Liberia, Madagascar, Malawi, Morocco, Nepal, Nicaragua, Uzbekistan, Papua New Guinea, Paraguay, Lao People's Democratic Republic, Dominican Republic, Democratic People's Republic of Korea, United Republic of Tanzania, Rwanda, Senegal, Sierra Leone, Sri Lanka, Tajikistan, Chad, Togo, Tunisia, Zambia, Zimbabwe
5 – brown	- Lower living standards and significant economic and political challenges - Demand for poultry meat may be limited due to economic constraints and food security issues - Expansion into these markets can be risky and would require a specific strategy focused on local development and product accessibility	Afghanistan, Bangladesh, Burkina Faso, Burundi, Cameroon, Ethiopia, Iraq, Kenya, Libya, Mali, Mozambique, Myanmar, Niger, Nigeria, Uganda, Central African Republic, Democratic Republic of the Congo, Sudan

7. Target Countries

• Red Cluster

These countries present an interesting potential for future expansion of our poultry products, but access to these markets is challenging due to strong competition.

• Orange Cluster

• Green Cluster

These countries are selected due to their high to medium averages and seem conducive to good market penetration.

Orange Cluster

‘Saudi Arabia’, ‘Austria’, ‘Belarus’, ‘Belgium’, ‘Bolivia (Plurinational State of)’, ‘Bulgaria’, ‘Chile’, ‘China - Hong Kong SAR’, ‘China, Taiwan Province’, ‘Costa Rica’, ‘Denmark’, ‘United Arab Emirates’, ‘Finland’, ‘Greece’, ‘Hungary’, ‘Israel’, ‘Kazakhstan’, ‘Malaysia’, ‘Norway’, ‘New Zealand’, ‘Portugal’, ‘Republic of Korea’, ‘Romania’, ‘Serbia’, ‘Slovakia’, ‘Sweden’, ‘Switzerland’, ‘Czech Republic’.

Green Cluster

South Africa, Colombia, Egypt, Iran (Islamic Republic of), Pakistan, Peru, Philippines, Thailand, Turkey, Ukraine, Vietnam