More and more we are all coming to realize that everything around us is data—it’s up to us to learn and determine how to leverage these data.
The growing capability in the graph data base and analytics technology space means that we have more opportunities than ever to harness the data around us, but also ever-increasing challenges to avoid the pitfalls of Maslow’s Hammer
When we take on a new project, we make sure to sit down with stakeholders to ensure we can achieve the goals of the project:
Addressing not just what’s wanted, but what’s needed—this helps ensure that we are using the right tools for the job
All this to say that we’ve seen a steady increase in clients interested in leveraging the physical data around us—from those in telecommunications and energy industries to logistics and even human resources!
We essentially approach workflows holistically, but comprehensively across areas relevant to any project:
Software and vendor selection
System automation and design
App development
Business Analysis
Product Management
Sustainment
While we don’t have the liberty of openly discussing the specifics of the work in these partnerships, we would like to demonstrate a project, couched in our general workflow, in which we take a complex problem with which many of us are familiar, leverage graph technology to facilitate intuitive analyses by modeling a physical system and taking all of this to the next level with cutting edge, intuitive visualization
One question that we were interested in exploring was if covid would spread through ventilation systems. So we thought it might make a good basis for a demo of graph + visualization.
This also helps illustrated work we are doing on actual projects .
The challenge of this demonstration on its face is concrete:
A hypothetical client representing a school board came to us with a request to help assess risk in the HVAC system of on of their high schools.
Ensuring HVAC is in top shape is important at any time, but as we know, this profile has been raised thanks to COVID
This leads to a couple of questions/asks:
How can we proactively address possible risks in our HVAC with respect to the spread of contaminants like COVID? Where are the riskiest chokepoints or junctions from which COVID could spread easily?
If we do experience contamination, is it possible to determine the source of any contaminants? Or if we know the source, can we determine the extent of possible spread and contamination?
Is it possible to visualize, or ‘inspect’ our HVAC and any analyses we do on it in an intuitive way to help speed up reactions?
Following consultation with the stakeholders to define the scope: both the challenges to address and the availability of data to do so, we planned how to accomplish this task:
The board had technical specs and a design model of the building, including HVAC, in REVIT
Utilizing graph database technology provides an intuitive way to model complex physical systems featuring lots of interconnectivity and/or dependencies:
Buildings/architecture
Transportation and logistics
Energy and natural resource systems like pipelines or processing facilities
The list goes on!
Finally, to facilitate and expedite data-driven decision-making, we opted to bring together the physical model and analytics together in UNITY: a widely-used 3D modeling platform
Why 3D model?
Communicating results is just as—if not often more—important than the analytics or data science that power any deliverables
Visualizing findings in the abstract using tables, bar graphs, pie charts and the like is minimally viable, but why not make it truly intuitive?
Why not visualize the graph and any analytics embedded in our 3D model—that is, visualize our analysis in the building itself, as if we were conducting virtual inspections?
Why be satisfied by being told the what when we can instead show both what and where?
Before we go on…a few words of thanks
First step is to bring the REVIT model into Graph
MIMO TO LEAD DEMO PART 1
This slide begins Demo part 2 (stuff in neo4j)
PETE TO LEAD DEMO PART 2
With the model in graph, we can traverse the HVAC system in graph as tough we were “walking the ducts” from rooms or spaces to terminals to ducts, equipment, duct transitions
[PETE to RUN ERIC’s DEMO QUERY]
We can leverage GDS to assess risky hotspots—either places that could be spreader sources or chokepoints
Recommend monitoring, upgrading components and/or filtration in these areas
Similar to how we might try to find bottlenecks when dealing with transportation/logistics
We can further leverage GDS to conduct probabilistic contact-tracing, either forward or backward
Proactively simulate forward tracing – you know where COVID-cough took place
Reactively simulate backward tracing – you know where COVID appeared, but not where it originated
[PETE to DEMO IN NEO4J BROWSER FOR GDS, run through queries/algo results]
As we can see, tables and graphy visualizations take us a good lot of the way there, but how can we make our traversals and analyses more intuitive, and expedite the path-to-value in decision-making processes?
Which is easier: trying to find something that showed up in a list/table, or getting shown exactly where that something is on a scaled map/visual representation?
MIMO TO LEAD DEMO PART 3
Information management throughout the organization:
What needs to be seen: when, where, why, and how?