An enterprise mobile and desktop search app for Providence St. Joseph Health that gives caregivers a single place to discover, collect, compare, and access business intelligence data, analytics, and applications from numerous isolated source systems.
In 2014, several years of business consolidations and mergers had left Providence data assets spread out across a wide variety and number of source systems and vendor platforms. Looking to promote data unanimity through through visibility, I cobbled together an open source search application on five retired desktop PC's under my desk.
Capital funds from a Data Governance initiative legitimized the app, and allowed me to improve the app's user engagement features like natural language queries, search suggestions, automatic faceting, result boosting, bookmarks, saved searches, comments, and "likes".
In 2015, Providence evaluated several commercial data catalog vendors and was unable to find a good fit. Their products were too expensive, too time-consuming to manage, too legally constraining, or not readily adaptable to ad hoc user feature requests. In the end, Providence leadership decided to operationalize my app and make it available to all internal users for free.
They called the app 'myHIway' based on the self-service nature of the application and its association with the Healthcare Intelligence (HI) team. From a user's point of view, the name meant "I can have 'HI' data ‘my way’ or "I can find data assets ‘my way’ using an 'HI' built application." From a business strategy perspective the name implied "health care transformation through faster access to data."
In 2016, I released the first version of myHIway on Providence datacenter VM servers. As hoped, the product accelerated report consolidation and data quality initiatives by exposing over 40k business intelligence apps, reports, metrics, datasets, analytics, dictionaries, standards, data models, glossaries, training materials, and data science references from over 30 internal and external data and content sources.
In 2020, with a stable user base of over 14,000 enterprise caregivers, the app's value and reputation was good enough to justify investments that improved the app's user interface, added new features, and metamorphized it into a fully opensource microservices-based on-premise Kubenetes solution.
In 2022, we legitimized visions for future enhancments by engineering the app as a single sign-on PaaS solution on Azure Kubernetes Services (AKS). Through streamlining and automation, the team was able to more readily take on complex feature requests.
UI Features
- Asset Galleries
- Search - NLP, auto-suggest, result boosting, saving, facets, and filters.
- Bookmarks
- Saved Searches
- Recently Visited
- Comments
- Access Request Forms
- Personal Preferences
- Single Sign On
- System Alerts Banner
- WordPress Blog - for research, data news, and general data literacy posts
Platform Features
- 8 Data Ingesters
- Automated Database Backup
- Service Logging
- Click Tracking
- JWT Validation
- SolrCloud
Responsible for product design, architecture, and full stack engineering from proof-of-concept through current releases. Created brand logo and original graphics/theme elements.
2014 to 2016
JSP/LESS web application built on Twigkit framework running on Apache HTTPD and Tomcat servers connecting to Apache Solr with data modeled using SmartLogic and ingested from Enterprise Data Warehouse and MySQL databases on Desktop PC's pirated from empty office cubes.
2017 to 2020
Angular2/LESS web application built on Lucidworks App Studio framework running on Apache HTTPD and Tomcat servers connecting to a Lucidworks Fusion search engine with data ingested from Enterprise Data Warehouse and MariaDB databases on Providence Cloud VM servers.
2021 to 2022
Fully opensource Angular9/SCSS web application, Apache Solr cluster, and numerous dockerized Python microservices for logging, token validation, and data ingestion from SAP Business Objects, Epic Clarity, Snowflake, SharePoint, Tableau, PowerBI and MariaDB datasets. Deployed from Azure DevOps Repos and Azure Container Registry to an on-premise Kubernetes cluster on Providence Datacenter Cloud virtual machines.
2022 to present
Packaged by Helm and deployed to Azure Kubernetes Services as a scalable, fully self-contained PaaS solution in a single namespace made up of 45+ pods that pulls its secrets from an Azure Key Vault and automatically backs up three databases to an Azure Fileshare storage account.