RTCNA2016 Recap - "Computational Design for the 99%"

Several weeks ago I had the privilege of presenting at Revit Technology Conference – North America 2016. My presentation frequently repeated the phrase “Because Nobody Went to Architecture School to…” We have all been there at some point in our career – continuously repeating the same manual alteration to a Revit model, changing parameter information one click-at-a-time, or performing tedious data entry for hours on end – these are the moments when you wonder if the practice of architecture is not exactly what you dreamed about in architecture school. For all the advancements that BIM has introduced to the AEC industry, production validation, and maintaining uniformity of the information are still difficult undertakings. Tasks that require hours and days of individual modifications are not always professionally rewarding and monopolize time that can better be spent on the overall quality of the design and documentation. I often tell colleagues that if you find yourself asking the question, “There has to be a more efficient way to do this”, chances are good that Dynamo can help.

I did not come from a computer programming background but instead began teaching myself Dynamo to address specific problems frequently encountered in Revit. After achieving a basic understanding of how Dynamo works, I was able to investigate tasks of increasing complexity that began with simple changes to the model and evolved to automating entire processes. As my Dynamo experience continued to grow I began exploring ways that Revit could interact with other software platforms and how data could be manipulated and visualized. My skillset eventually evolved to where I understood more advanced concepts of geometry and parametricism for design but this was all built on the foundational knowledge acquired from researching daily production tasks.

In my presentation I preceded to share a sample of workflows that respond to specific challenges encountered on projects and tell the story of tedious task automation and process improvement for architectural practice. A highlight was the opportunity to collect data on a very large healthcare project that I developed into a workflow for tracking Revit model metrics. The goal was to look for correlations between various model metrics and how long it takes to sync or open the model — one of the most significant factors of workshared projects because the extra seconds and minutes it takes to sync on a slow model multiplied by all the users on the project adds up to many hours of lost productivity over the course of the project. Dynamo is used to track the overall size of the .RVT file, query and count various elements and categories, parse the Warnings export file, then export all the information to an Excel file. In addition to collecting these general model metrics, the Dynamo task updated two additional spreadsheets with every warning in the model over time and every placed family in the model over time. All three of these of these spreadsheets were linked into Microsoft PowerBI along with data from imaginIT Clarity’s Model Metrics tool, which tracks the time it takes to open the model over time. Over the course of three months, I ran this Dynamo definition on a daily basis for a total of 68 exports.

The final takeaway will not be a surprise to those who are familiar with Revit model performance… the data revealed that Auditing and Compacting the model as well as Purging Unused Families had the most overall impact on the time it takes to open and sync the model. Although this may not be a significant breakthrough, these real-time analysis tools help monitor the health of the model and indicate when may be the best time to intervene.

The last step was to find an easy way to communicate the status of the model to the production team. Since it is the responsibility of the Model Lead on the project to audit the Central file, Warnings are the only characteristic that individual team members have the opportunity to impact. The project from which this data was collected happens to be a children’s hospital so we placed an image of a Minion on the Message Board with a visibility parameter tied to the number of Warnings. The final Dynamo task overwrites the Warning count parameter in the Revit model and the Minion changes accordingly. Now the team is aware that when they open the model at the beginning of a workday, if the Minion is purple the Warnings have exceeded 400 and some time needs to be set aside to resolve.

In the end RTC was an excellent experience. I thoroughly enjoyed sharing my perspective and bonding with my fellow colleagues from all over the world.

Special thanks to everyone who helped contribute to my work:
RTC & Committee
Shepley Bulfinch
Jim Martin
Jim Chambers
Jessica Purcell
Christina Tully
Margaret Gammill
PJ Centofanti
Jamie Farrell

Stairway to... Mt. Washington

Every March a group of Shepley Bulfinch employees (and friends) makes an excursion to the summit of Mount Washington. Situated in the the Presidential Range of the White Mountains in New Hampshire, Mount Washington is notorious for unexpected weather changes. Average temperatures for the month of March are a high of 20 degrees and a low of 5, with sustained wind gusts of up to 60 mph at the summit that can easily create a subzero windchill. Hiking at the tail end of Winter often means that crampons, ice axes, interchangeable layers, and many pairs of dry gloves are required equipment.

The hike itself lasts approximately 7-9 hours, covering 10.5 miles and nearly 4,250 vertical feet. After many months of hibernation during the blustery Boston Winter, serious training is necessary to prepare for the quad-burning ascent. Beginning 6-8 weeks prior to the climb, individuals begin hiking the egress staircase of the office building with reams of paper in their packs to simulate the weight of equipment and provisions carried on the day of the climb. This year everyone began recording their training sessions to stir up a little friendly competition and here are the results...

There are 16 floors in the office building and an approximated 736 average steps per lap (all the way up and down once) after accounting for the stairs, landings, intermittent stretch breaks, etc. Therefore it is estimated that the entire team collectively climbed roughly 349 laps, 11,168 floors, and took 257,269 total steps.

It looks like Monday, Wednesday, and Friday were overwhelmingly popular for making the time to fit in stair training around busy work schedules and other committments.

Here is the breakdown of number of laps per person:

...And an analysis of the average time per lap:
Notice that as weight and amount of laps increased, overall time would often slow down significantly.

We lucked out this year with sunny, warm weather for the entire hike. Everyone summited successfully and it was another fun-filled team experience.

Special thanks to our fearless guide David Meek who year-after-year invests hours of preparation and tireless effort to ensure a safe journey for all!


Image courtesy of Hung Pham

Image courtesy of Hung Pham

Canstruction is an international competition where “cansculptures” are created using cans of food, which are later donated to local hunger relief organizations. This year, the Houston office of Shepley Bulfinch assembled a team to take on the challenge.

After coming up with a design concept, the task of estimating the number of cans and positioning for structural stability is daunting. Since cans of food are a modular unit with standardized dimensions, this is a perfect opportunity for parametric design using Dynamo.

The first step was to figure out how to populate cans along a surface. Following a visit to a local grocery store for research, the height and diameter of a chosen can were entered into Dynamo to establish the base module. An undulating vertical cylinder was created to a height of 6 Feet and cross-sectional circumference curves were cut based the height of the can.

Points were then placed along the horizontal curves at a repeated distance of the width of the can. In order to ensure structural stability, the last step was to offset the location of the points along every other curve by the distance of half a can width so that they would be perfectly staggered to land at the point of intersection of two cans on the level below.

For the final design concept, the character Blanky from the animated classic The Brave Little Toaster was chose to help convey the message “Stitching Away Hunger!”

The initial 3D object to visualize the Blanky’s body was created using NURBS in Rhinoceros and then imported into Dynamo using the Mantis Shrimp package from Archi-lab . Initial attempts at generating curves from the Rhino surface resulted in a significant loss of definition of features due to the NURBS curves in Dynamo rounding off sharp corners between points.

It was determined that a better approach would be to model the concentric curves, similar to a digital topography model. The geometry was transitioned to Revit where contour lines were generated at a vertical spacing of the height the can. As splines in Revit, the contour lines could be easily adjusted to maintain the precision of the shape’s features.

Dynamo facilitated a seamless iterative design process as the contour curves were adjusted in Revit then queried in the definition to populate with cans. Upon visual inspection of the resulting 3D geometry in Dynamo, further adjustments could be made to the curves in Revit to perfect the overall shape.


Cans tend to become unstable after a certain number of stacked layers so a common practice is to add a thin layer of supporting material at regular intervals to provide a firm horizontal surface. The team chose 1/16” medium-density fiberboard (MDF) as their horizontal reinforcement layer at every level of cans and added supplemental contour curves in Revit to simulate the 1/16” spacing. The curves were picked up in Dynamo and populated with an extruded surface to and visualize the addition of supports, as well as calculate the increased overall height.

To emphasize the distinction between the character’s cape and the void below, the decision was made to use taller cans of a different color for the lower central portion of the can sculpture. In order to account for this design alteration, the contour curves were strategically split in Rhino and – at areas to be substituted with larger cans – contour lines were removed to account for the increased height.


One of the most significant benefits of designing a Canstruction sculpture with computation is the ability to perform an instantaneous takeoff of cans and materials. For this cansculpture, the final count came to 1763 short cans and 120 regular cans. In addition, there were 42 layers of support material, which ended up requiring [60] sheets of MDF. These numbers were critical for designing within budget and placing supply orders.

Image courtesy of Billi Jo Galow

Image courtesy of Billi Jo Galow


Linework from Dynamo can be pushed back into Rhino or exported as an SVG file. In preparation for build day, the assembly team printed out full-scale templates at each level of the cansculpture, which allowed them to quickly place cans on top of the template, eliminating most unforseen variances and the need for improvisation in the field.

Image courtesy of Billi Jo Galow

Image courtesy of Billi Jo Galow


Congratulations to the Shepley Bulfinch – Houston team for constructing an impressive cansculpture for a good cause. It was a pleasure to assist with the use of Dynamo and interoperability among several software platforms for simulation and delivery of their design. They demonstrated that visual programming is easy-to-learn and invaluable tool for integration into the design and documentation process, whether producing a sculpture made of cans or an entire building.

Claudia Ponce
Hung Pham
Julie Truong
Billi Jo Galow
Sandra Bauder
Stan Malinoski