Why We Love Healthcare In Graphs
While the front-end team has been busy putting together version 3 of our healthcare marketplace, the data science team has been hard at work on several things that will soon turn into new products. Today, I'd like to give you a sneak peek at one of these projects, one that we think will profoundly change the way you think about health data. We call it the PokitDok HealthGraph. Let's ring in the New Year with some data science!
Everyone’s been talking about Graph Theory, but what is it, exactly?
And we aren’t talking about bar graphs and pie charts.
Social networks have brought the world of graph theory to the forefront of conversation. Even though graph theory has been around since Euler solved the infamous Konigsberg bridge problem in the 1700’s, we can thank the current age of social networking for giving graph theory a modern revival.
At the very least, graph theory is the art of connecting the dots, kind of like those sweet pictures you drew as a kid. A bit more formally, graph theory studies relationships between people, places and/or things. Take any ol’ social network - Facebook, for example, uses a graph database to help people find friends and interests. In graph theory, we represent this type of information with nodes (dots) and edges (lines) where the nodes are people, places and/or things and the lines represent their relationship.
As data science enthusiasts, we were excited to work with a graph structure because it opened the door to being able to use a wide range of mathematical tricks and algorithms. When using a graph, patterns between various people, places, and things (the nodes) quickly start to emerge. For example, highly connected nodes can often times be considered more influential or important than nodes that have fewer connections. Also, nodes with similar or shared edges and connections can be used to identify trends and make recommendations through inferred correlations. A quick search on Amazon will show you how fast a personalized recommendation can be made after browsing only a few products.
These are just a few examples, but the takeaway is that a large assortment of metrics exist that can be used to make inferences about the people, places, and things - and how they relate to one another. Our HealthGraph, similarly, gives us the power to answer all of our traditional data science questions, like ones relating to provider statistics, while also inspiring new ways to examine the healthcare system, like inferring provider recommendations using graph traversals.
Ok, got it. But how is this related to healthcare?
To make a long story short: healthcare is about you and connecting you with quality care. When data scientists think of connecting things together, graphs are most often the direction we go.
At PokitDok, we like to look at your healthcare needs as a social network, aka: your personal HealthGraph. The HealthGraph is a network of doctors, other patients, insurance providers, common ailments and all of the potential connections between them.
We built our HealthGraph accordingly and pulled together all the healthcare data we could get our hands on - from the American Medical Association, Medicare Co-Occurrence data (check out our previous graph visualization on this data set), our own research, and beyond. By building a HealthGraph, we can ask questions and make discoveries that would be difficult, if not impossible, with other traditional databases. Let’s face it, traditional (or relational) databases make storing, querying and understanding wildly different datasets more difficult than it should be. Modeling the healthcare system with the PokitDok HealthGraph solves those issues and yields a wide variety of applications and possibilities for analytics, but most importantly, it gives you the ability to connect with the quality care you deserve.
A sneak peak of what's next...
To give you a sense of what one of our data sets look like, take a look at the -dare we say gorgeous- image below. This represents all of the providers within the state of Washington and the similar features they share. The doctors are represented by the green nodes, while the healthcare organizations are shown in bright pink. Spend some time and try to draw a few conclusions of your own until our next post when we take a deeper dive into the PokitDok HealthGraph and what it means for the future of healthcare.
- AWS Presents: Managing a Healthcare System with Blockchain - June 28, 2017
- DocRank: Computer Science Capstone Project with the College of Charleston - June 5, 2017
- Notes from the Field: Consensus 2017 - May 31, 2017
- Snake_Byte #28: Term x Document Matrices via Sklearn's DictVectorizer - April 21, 2017
- DC Blockchain Codeathon - April 11, 2017
- Snake_Bytes #5: Vector Distances - October 12, 2016
- TinkerPop and Titan from a Python State of Mind: A Tutorial via Docker - December 15, 2015
- Online Recommendation Systems: Getting Personal with Gremthon - Part 2 - April 14, 2015
- Online Recommendation Systems: What it Really Means to be 'Somebody Like You' - Part 1 - April 7, 2015
- The top 3 things we love most about the PokitDok HealthGraph - January 27, 2015
- Snake_Byte #35: Building Custom Slack Bots with Python - Mimic Your Coworkers - August 11, 2017
- Snake_Byte #14: Writing Made Easy - January 6, 2017
- Top 100K PokitDok Providers - November 30, 2015
- Visualizing Medicare Physician Co-Occurrence and Payment Data Geographically - Part 2 - July 14, 2015
- Visualizing Medicare Physician Co-Occurrence & Payment Data Geographically - Part 1 - July 9, 2015
- National Provider Data Revealed - A Choropleth Map Analysis - October 9, 2014