Module 3
Graphs, Graphs, more Graphs.
I am actually a big fan of graph theory. I first got into it when I was maybe around 14 years old and was writing a path finding algorithm for a simple video game I was creating. Graphs seem to represent a solution to many problems in the field of computer science. They are using them more and more. As the lecture suggested Google uses graphs to rank it's search results. Which makes alot of sense when you think about how widespread the internet really is. The lecture this week focused on all the different types of graphs and the various measurements of graphs and so-forth. Some of the measurements were simple, aka, how many nodes is this node connected to. Some of them were complex, like the number of paths that ran through this node to get to other nodes. It's amazing to me how simple and complex a graph can be. It makes sense we're learning more indepth for them in this class because I imagine a ton of business intelligence can be visualized and represented in graph form and it seems like we're also going to be doing a project involving graphs and linked-in so I am excited.
For my independent research this week I wanted to share something I'm already vaguely aware of but could be very useful if you haven't researched it before. I would highly recommend reading Kumar's guide to graph databases. No, it won't take you to white-castle, but instead it will teach you about all the great developments going on with graph databases. This article really just tickles the surface of graph databases and I would only recommend it as a reading to get your feet wet. In order to truly appreciate a graphdb you will have to play around with one. One key take-away from the article is that they are using graphdb to combat fraud, and also to power recommendation engines. The future is wild!
Sources:
Kumar, Vivek. “Graph Database: What Is It and Why It Matters for Businesses?” Analytics Insight, 25 Sept. 2020, www.analyticsinsight.net/graph-database-matters-businesses/.
Hi Todd, Thanks for sharing....Kumar's guide to graph database was a good read. Clearly, there is more to technological advancement that will necessitate answers! The article by Barabasi regarding computational science also alluded to this.
ReplyDeleteHi Todd, as I was reading Kumar's post I remembered that I have 2 books on graph databases still waiting to be read. One of those things one never seems to find time for.
ReplyDeleteNow that I gave some thought about it, I can't help to see some analogies between graph databases and the classical geodatabase model, which uses vertices and arcs (=edges) to represent tessellations (e.g. US counties) without redundant data (so one vertex represents that vertex for all the polygons that share it, and one arc represents that arc for all the polygons that share it). The short story is, it got me wonder if/how graph database could be used to optimize geodatabases. Geodatabases are currently well tuned thanks to spatial indexes (e.g. R-trees, quadtrees), but improving performance will always be welcome... :)
Hi Todd,
ReplyDeleteThank you for introducing the idea of graph databases! It got me thinking all the way back to our first assignment in the class which revolved around star schemas and dimensional modeling. I found the link when Kumar's article mentioned relational databases and graph nodes being faster joins. Could you simply say that a graph db is a dimensional model in a different form. I get the feeling that Kumar is hinting towards the usage of dimensional modeling with newer NoSQL based databases to get information quickly.
Thanks for sharing.
-Dustin