POC for Automating Forecasts for Revenue, EPS and P&L data

Problem

  • The client team has a manual process to perform Forecasts on Revenue, EPS and P&L data
  • Aim of the POC was to build on the manual top-side Revenue/EPS model and Lighthouse predictive analytics experiences to develop largely automated “Top-down” revenue, EPS and P&L forecasts

Solution

The custom-built forecasting and prediction PoC supported:

  • Data Transformation to convert the input data for forecasting
  • Time Series and Machine Learning Algorithms (Linear Regression, Arima, FBProphet and NeuralNet algorithms) using R and Python
  • Benchmarking of the AI-driven forecast against Manual forecasts to identify the best fit forecast to actuals
  • Exploring multiple ways of running forecasts using Python Apps, RShiny Apps, batch jobs.

Impact

01

The new algorithms were benchmarked against manual forecasts and found to be close to actuals

02

Ongoing experiments on additional datasets to validate the approach

Scroll to Top