Predictive Analysis for Chemical Delivery Frequencies and Demand

Project Description

We are a chemical distribution company focusing on the aquatic industry across the Gulf Coast and are looking to utilize our data to predict the most optimal delivery dates for our product. 55% of our sales come from customers that pay us a flat rate and in return we provide whatever quantity of chemicals is necessary.  Therefore we are in a position to determine when the best time to deliver would be (we aim to deliver when 80% of the product has been consumed).

Our customer groups include water parks, apartments, hotels, water treatment plants etc. Currently we have ~5500 active delivery points and would like to potentially use the historical delivery data, customer information, and outside factors (ex: weather) to predict when the next time a customer will need a delivery.  We have up to 4 years delivery data in our database (NetSuite) that can be exported to any type of file or integrated in many ways.

Currently we use a very basic r-language model that is only looking at historical data and is not extremely accurate so is more used to identify accounts that need to be looked into further.  We utilize cloud applications across our organization therefore the ideal solution will follow this strategy.

An example of data includes: delivery amount and date, pool volume, account type (apartment, hotel, etc), location, Apartment specific information including Occupancy, number of units).

In your proposal please provide total amount of hours you will require to complete project.

Thanks!