In this project, I tried to assist a manpower agency to identify which country is best fit for a specific client based on client specifications. I also wanted to give a business solution to increase their profits by showing them the countries that best return the highest profit.

First, I acquired a row and dirty data from the agency office that has been collected between the years 2014 and 2015. The first data came in three excel files and it was all associated with the countries Bangladesh, Philippines, and Sri Lanka. The data did not look like an easy and clean data to work with, which results in spending hours and hours of preprocessing which includes cleaning, extracting features, etc. Then, I combined all of the three datasets into one, so that I can perform some modelling and visualizations. I used this combined data to perform classification modelling which means that I wanted to predict the probability that a country would be the best choice for a client based on his/her specification. I also perform clustering modelling so that I can study the behaviour of clients. In the end, I ended up with a good accuracy model that predicts the probability of the best country that the client would like based on his/her specifications.

Second, I spent a lot of time looking for different datasets on the agency online system, so that it would support my modelling process or help me to come up with another solution based on the first prediction. I found a really useful data that shows the price of recruitment, cost of the supplier, and the profit. Thus, I used this data to show the profit among countries after I predict the country in the first part so that the agency can use the model and the data analysis to the benefit of the business. In conclusion, service providers can benefit a lot from the model and the data analysis.

Third, I broke my project process into three files. dirty_data where I got my data as a raw data and then I cleaned it. In the second file 'clean_data', I performed data visualization and drumming my data to prepare it for modelling. In this file, I performed a classification and clustering models. I the third file named 'profits_file', I visualized data after cleaning so that we can use them in the prediction file 'clean_data'.

To sum up, I believed that I did good work achieving a better accuracy in my classification model that would be beneficial for the business. I also believed that my data analysis and visualizations would assist the employer to make better decisions. For future work, I am planning to work hand to hand with my uncle to build a system that collects data properly so that we can make better-looking visualization which would result in making better business decisions