Process flow

The manual planning of project resources often lacks precision due to complex schedules, varying project classifications, and dynamic progress rates. This AI-driven solution enables proactive, data-informed planning using historical project data.

The solution consists of these components:

Preprocessing
  • Reads project metadata such as address, business sector, acquiring method, financing method, and geographical area.
  • Preprocesses the raw data coming from datalake and identifies comparable closed projects for training. Semifinished and unfinished projects for inference.
Activity progress prediction leveraging Random Forest algorithm
  • Uses the scheduled and actual progress patterns from historical projects.
  • Aligns progress predictions with the Work Breakdown Structure (WBS).
  • Predicts percentage completion for future periods.
Cost object quantities prediction leveraging XGBoost algorithm
  • Uses predicted physical progress and historical consumption trends.
  • Outputs forecasts by activity, cost type, cost object, and period.
  • Forecasts actual quantity consumption by period.
Widgets Display
  • Displays forecast results through configurable widgets.
  • Enables easy monitoring of fulfillment trends and risks.

The solution processes the data through ION Data Fabric, organizes the data using Data Catalog, and utilizes Infor AI for predictive modeling. Results are then displayed via Widgets for actionable insights.

Data flow

The data flows through these components:

  1. ION Data Fabric: Collects and processes data from Infor LN CloudSuite.
  2. Data Catalog: Organizes contract and historical transaction data for analysis.
  3. Infor AI: Runs predictive models and generates fulfillment forecasts.
  4. Widgets and Dashboards: Displays results for business decision-making.