CM4050-DRP000779

Faculty: Raghuraj Pandya

Forecast steel price using machine learning.

Construction industry is challenged with low margins and ripe for disruption. Given India’s infrastructure development steel demand is significant. In addition to supply, economics, raw material costs and export demand significantly impact on the market price. Ability to better forecast input costs like that of steel could help with timely procurement and cost management for construction contractors and clients. This research is building on further from last year’s DRP research where time series tools were used to forecast steel price 12 months ahead, using Prophet time series ML model. Further it was identified that raw material price namely crude and metal prices have substantial influence on the price of steel. Implies the cost of steel is largely dependent on input cost, while that of cement was identified to be driven more by demand. This could be either due to the competition in the industry or macroeconomics or construction specific use case. Proposed research will look to improve on the research last year and build-in more advanced analytics/machine learning model or include additional external predictive factors or study detailed dynamics of steel price and the input/demand mechanics. This is an opportunity for students with domain experience to collate data for this construction challenge. And under guidance from a lead data scientist to develop a solution that helps construction industry players realize value from the forecast. And in the process student will gain data science skills and more importantly soft skills relevant in the current industry.