US Air Force | Automated & Augmented Analytics for CMXG Jobs & Backorders


The Air Force is enhancing the analytics required to effectively manage corrective and preventative maintenance for aircraft and weapons systems. The existing process requires data to be aggregated from multiple disparate sources and involves data preparation, synthesis, analysis, and reporting. However, this process is time-consuming and cumbersome connected spreadsheets diminish leadership’s ability to understand drivers and intervene to solve problems at a granular level.

Shipcom was assigned the challenge of re-envisioning and automating the Daily Maintenance Output Report (DMOR) which communicates performance analysis to the Air Force maintenance operations leaders. This task includes enhancing the monthly presentation of key performance indicators, job completion rates, supply constraints analysis, the calculation of flow days and flow days variances, the impact of backorders on MICAP, squadron performance variances, and the overall impact of these metrics on Mission Capability (MC). Analysis of these variables by Air Force leadership will increase their ability to correct problems quickly and reduce risk.

The cornerstone of the solution is a semantic data fabric, which contains a knowledge graph of all relevant maintenance data items. The data in the fabric fuel the analytics and visualizations that deliver diagnostic insights into maintenance operations. The analytics generated by the system is presented in Shipcom’s Catamaran, an AI-infused applications suite run in an IL5 secure environment.


1. RE-IMAGINING THE REPORTING WORKFLOW: Shipcom’s solution significantly improves the time required to transform the maintenance data for analysis by eliminating manual data collection and data integration work steps., The resulting workflow automates data preparation by deriving new features from original variables, applying conditional logic, and reshaping tables when appropriate. The workflow transports data through a series of value-adding transformations and delivers the data to the analytic programs that derive insight from the data.
2. DATA FLOWS: The data required to analyze maintenance operations are drawn from several enterprise databases. The dataset is assembled from diverse data sources, including Lean Depot Management System (LDMS), Logistics, Installations, and Mission Support (LIMS-EV), Impresa – to focus on open, closed, and backorder jobs to provide insightful analysis. These data are stored in a semantically aligned knowledge graph which provides a method of storing and organizing data that emphasizes relationships between data items and allows for adding and modifying entity types and relationships without extensive database restructuring. A knowledge graph is a database that employs a graph-structured data model, facilitating the seamless integration and interconnection of data. It can manage a broad range of structured data types, encompassing data that describe events, logical conditions, and objects, making this a logical choice for storing interconnected maintenance data. A knowledge graph improves the analytic output by creating relationships among data items, allowing inferences drawn from linked data. Storing data in a graph rather than a relational database facilitates the semantic description of data items, yielding a deeper understanding of the data. This capability allows the extraction of new insights that are not explicitly captured in the raw data.

3. VISUALIZATIONS: Interactive visualizations of maintenance data give decision-makers deep insight into the factors that drive the performance of maintenance operations. The dashboards enable drilling down into each visualization to understand the meaning of the data.

DMOR – CMXG Jobs: The first dashboard in the application presents an overview of the maintenance operation’s key performance indicators. These data allow users to examine variances to identify problem areas and formulate strategies to fix them. Furthermore, the application provides functionality for users to slice data by Squadron, Job Type, Open and Closed Conditions, and Flow Days. This more granular view of performance-impacting factors enhances the accuracy of problem-solving. This also helps identify flow day anomalies and uses AI to provide possible explanations for the anomalies.

Backorder Analysis: The Backorder Exploratory Analysis allows for exploring backorders by various key categorical fields (dimensions) providing a hierarchical view of the data and the factors contributing to a result. In this chart, the Backorder quantity helps identify variations in backorders by Source of Supply, the Shop doing the work, a Noun phrase that describes each part, and the NIIN, the National Item Identification Number used by the DOD to uniquely identify a part so that targeted solutions can be developed. The backorder influencer visual uses AI to identify key influencers that result in an increase in backorders.

Flow Days Analysis: Flow Days measure the cumulative time for a maintenance job to be completed, including work and wait time. This dashboard illustrates the difference in actual Flow Days vs. expected, by Squadron and Job Type and how the variances change over time. The ‘Ask a question’ feature in the top-right allows for AI-driven natural language processing (NLP) to ask direct questions about your data and visualize it, in addition to what is already presented.


The automation of the Daily Maintenance Operations Report transforms the process of creating and analyzing management reports, which will enhance the Air Force leadership’s understanding of the factors driving the performance of maintenance operations. Rationalizing the process lowers costs and increases the efficacy of the reports. The analytics and visualizations delivered in Shipcom’s Catamaran will significantly streamline report preparation and analytics, boosting leadership’s confidence in the data used to solve critical maintenance issues that impact mission capability.

Detailed Insights on

manufacturing jobs over 12 months
0 +

Drill-Down Analysis for

0 +

Data Pipeline Automation can save approximately

$ 0 /yr

AI augmented analytics can result in reduction of flow days by

0 %

Recent Articles