Create a Machine Learning based cost estimation tool for municipal construction projects.
The process of generating budget and timeline estimates for commercial and municipal construction projects is slow and error prone. We were asked to create a prototype of a system to accept project specifications and requirements, and generate accurate estimates by applying Machine Learning to analyze the actual budgets and timelines of previous projects to better understand and predict the budgets and timelines of similar projects.
We began with in-depth interviews of experienced estimators to define the tasks they need to accomplish to create an estimate. Then we looked at known databases of construction material and labor costs to derive a model which the ML algorithm could use as reference for generating cost estimates based on input from the estimators. Finally we created an online Functional Prototype to validate the workflows and simulated output of the application.
The startup was able to use the Prototype as a demonstration to potential investors and successfully raise enough funding to begin development of the application.