Project: Development of a mathematical model to forecast costs and revenues of internet provider
Problem: Our client ASARTA creates infrastructure of high-speed internet in the cottage towns on the territory of Russia. It needed to calculate the amount of the costs (8 different types of costs) that will be needed to provide 130 villages with Internet, the amount of funds required to support infrastructure, as well as income, which would bring the settlements.
Solution: We have developed a mathematical model that predicts:
- Infrastructure creation expenditures based on the available data (about 40 different items)
- Maintenance expenditures
- Villages commissioning rate (several villages per month)
- Planning Horizon – 3 years
Results: A model of costs and revenues prediction for 130 villages was made.
Technologies and methods: Python Libraries (Pandas, Numpy, Sklearn, Scipy), Methods (Factor analysis, Correlation analysis, Spline interpolation, Regression methods like: [ARDRegression, BayesianRidge, ElasticNetCV,…]).