Price forecast of building materials
About project
The model is constantly running. The main effect was achieved during a period of price volatility and made it possible to make a number of decisions on the purchase of building materials in advance.
The company needs to conduct a comprehensive time series analysis of the cost of building materials.
A task
Bisiness effect
Part of the position of building materials makes up a significant part of the cost. The key factors influencing price spikes are identified, many of which are macroeconomic. Based on these data, a model is built.
Solution
Stages of development
Technology and innovation
1.
Finding external data
3.
Model building
2.
Data parsing
4.
Model Implementation
5.
Model monitoring
In time series forecasting, the Greykite library from LinkedIn was used. Also in support was tested by ETNA from Tinkoff. The Airflow library set was used to plan and monitor workflows
Project team
Michael V.
Developer
Alexander S.
Analyst
Areas of use
This technology can be used by construction companies when deciding on the purchase of building materials for the future. The model works on an ongoing basis and reports are sent regularly: this is especially useful during periods of price volatility.
