Data Science & Advanced Analytics

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Good afternoon! Daniel, I'm in a very different project with Petrobras (Brazil) where we had to implement low code tools and follow AWP Data Requirements Implementation.

https://www.construction-institute.org/awp-data-requirements-implementation-guideline-version-1-2

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Thanks Henrique.  One of our current challenges is classifying historical data that isn't necessarily packaged in a useful way.

Good afternoon! Daniel, I'm in a very different project with Petrobras (Brazil) where we had to implement low code tools and follow AWP Data Requirements Implementation.

https://www.construction-institute.org/awp-data-requirements-implementation-guideline-version-1-2

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Hi Daniel,

what kind of info do you need, are you thinking on how to build the info pipeline, setting up the data, setting up the servers, doing the data cleaning or just in general the whole process?

we did implemented this in my last company and now i am repeating the process with a whole new dataset so happy to chat up with you if you want

Thanks Henrique.  One of our current challenges is classifying historical data that isn't necessarily packaged in a useful way.

Good afternoon! Daniel, I'm in a very different project with Petrobras (Brazil) where we had to implement low code tools and follow AWP Data Requirements Implementation.

https://www.construction-institute.org/awp-data-requirements-implementation-guideline-version-1-2

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

My initial inquiry was more around the topic of standing up a team from a personnel standpoint - structure, personnel attributes, operational focus vs. IT focus, where it fits in the company operating structure, funding, justifying the cost, etc.  

Less interested in the details of process - well on our way in terms of classifying historical data, mainly using machine learning.

------------------------------
Daniel Gilmour PSP
Chair, AACE Planning & Scheduling Subcommittee
Director of Planning & Scheduling
Brasfield & Gorrie LLC
Atlanta GA
(908) 489-8819
dgilmour@brasfieldgorrie.com

Original Message:
Sent: 05-29-2024 08:10 AM
From: Luis Antonio Garcia Loera
Subject: Building a Data Analytics Team

Hi Daniel,

what kind of info do you need, are you thinking on how to build the info pipeline, setting up the data, setting up the servers, doing the data cleaning or just in general the whole process?

we did implemented this in my last company and now i am repeating the process with a whole new dataset so happy to chat up with you if you want

------------------------------
Luis Garcia Loera CEP, MBA,PMP
Head of Cost Engineering
Germany

Original Message:
Sent: 05-28-2024 03:17 PM
From: Daniel Gilmour
Subject: Building a Data Analytics Team

Thanks Henrique.  One of our current challenges is classifying historical data that isn't necessarily packaged in a useful way.

------------------------------
Daniel Gilmour PSP
Chair, AACE Planning & Scheduling Subcommittee
Director of Planning & Scheduling
Brasfield & Gorrie LLC
Atlanta GA
(908) 489-8819
dgilmour@brasfieldgorrie.com

Original Message:
Sent: 05-28-2024 01:57 PM
From: Henrique Diniz S. Silva
Subject: Building a Data Analytics Team

Good afternoon! Daniel, I'm in a very different project with Petrobras (Brazil) where we had to implement low code tools and follow AWP Data Requirements Implementation.

https://www.construction-institute.org/awp-data-requirements-implementation-guideline-version-1-2

------------------------------
Henrique Silva
Jaboatão dos Guarara
81994764004
h.diniz@outlook.com

Original Message:
Sent: 05-28-2024 12:57 PM
From: Daniel P Gilmour
Subject: Building a Data Analytics Team

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

------------------------------
Daniel Gilmour PSP
Chair, AACE Planning & Scheduling Subcommittee
Director of Planning & Scheduling
Brasfield & Gorrie LLC
Atlanta GA
(908) 489-8819
dgilmour@brasfieldgorrie.com
------------------------------

My initial inquiry was more around the topic of standing up a team from a personnel standpoint - structure, personnel attributes, operational focus vs. IT focus, where it fits in the company operating structure, funding, justifying the cost, etc.  

Less interested in the details of process - well on our way in terms of classifying historical data, mainly using machine learning.

Hi Daniel,

what kind of info do you need, are you thinking on how to build the info pipeline, setting up the data, setting up the servers, doing the data cleaning or just in general the whole process?

we did implemented this in my last company and now i am repeating the process with a whole new dataset so happy to chat up with you if you want

Thanks Henrique.  One of our current challenges is classifying historical data that isn't necessarily packaged in a useful way.

Good afternoon! Daniel, I'm in a very different project with Petrobras (Brazil) where we had to implement low code tools and follow AWP Data Requirements Implementation.

https://www.construction-institute.org/awp-data-requirements-implementation-guideline-version-1-2

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

I started the Planning Analytics Department on my own and have grown it into a team of four. Due to our success, I've been asked to expand our focus to include Cost & Safety and to lead the Operational Excellence Analytics Department. But enough about my achievements. :-)

Here are some answers that might help if I understand your question correctly. Feel free to reach out for a more detailed discussion.

Selling the value and service to management is crucial, especially since these positions are typically non-chargeable, adding extra pressure to demonstrate value. I was fortunate to have strong support from the planning and scheduling head, who was a visionary and believed in the power of analytics. Together, we established the analytics framework with a vision of predicting schedule risks across the portfolio and enhancing scheduling through best practices, technology, and process improvements.

thanks !

My initial inquiry was more around the topic of standing up a team from a personnel standpoint - structure, personnel attributes, operational focus vs. IT focus, where it fits in the company operating structure, funding, justifying the cost, etc.  

Less interested in the details of process - well on our way in terms of classifying historical data, mainly using machine learning.

Hi Daniel,

what kind of info do you need, are you thinking on how to build the info pipeline, setting up the data, setting up the servers, doing the data cleaning or just in general the whole process?

we did implemented this in my last company and now i am repeating the process with a whole new dataset so happy to chat up with you if you want

Thanks Henrique.  One of our current challenges is classifying historical data that isn't necessarily packaged in a useful way.

Good afternoon! Daniel, I'm in a very different project with Petrobras (Brazil) where we had to implement low code tools and follow AWP Data Requirements Implementation.

https://www.construction-institute.org/awp-data-requirements-implementation-guideline-version-1-2

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Daniel, I have been involved in multiple database development initiatives from the 1990s (yeah, I am that old) through just last year with a major owner. Structuring, capturing, cleaning data is table stakes for analytics (which by and large calls for structured data). Well covered in RP 114R-20. Serious firms have to set up dedicated resources for this, plus train staff down the line to play their part. The next step is the analytics itself which today means machine learning (old school regression up to now). Some analytics can be done inhouse, but one is likely going to want to leverage one of the many startups with high powered skills, and sometimes industry data (or models from it) that can be leveraged. The trick is how to marry analytics expertise (e.g., new grads with data science skill) with domain knowledge like estimating, scheduling, control and risk quantification who have never done a regression analysis. Most initiatives fall apart after a year or two when management finds out it cost money and getting really useful results is not immediate. 

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Dan - Let's chat about a historical database at the Conex, but here are the high-level thoughts based on two different firms in which I helped build a scheduling historical database:

1. Our reasoning to the executives to spend time on this important tasks was three fold:
   1. There were plenty of times when our preconstruction was asked the cost and schedule when they simply were given a SF and floors on a building. They were able to get to the cost pretty easily based on the $/SF, but the schedule duration was difficult. Based on a database, they can easily at least give a range based on last 5 similar projects instead of making up the durations completely. 
   2. We also wanted to stop fighting between the scheduling department and the preconstruction department at the time of proposals on what's a reasonable duration, if we completed a project in the same duration, if there was some overtime used, or if it was somehow different from the current proposal. 
   3. Also, track if there has been a pattern of delays for similar project types in the last few years and if we need to update our proposal strategy when bidding on these jobs. 
2. One challenge we had from the get-go was trying to use the preconstruction database and somehow use it for a scheduling historical database. In both cases, we created our separate fields and executed separate databases because the information was so different for cost vs. schedule. 
3. We spent a lot of initial time figuring out what we needed to do with the database and creating fields accordingly. At this point, you can also decide if you want to use software or self-build an in-house database based on your needs. 
4. We ended up hiring an intern or entry-level professional to outsource our data entry tasks initially. We were less worried about getting all the data in the database and more worried about inputting good-quality data. For example, we focused more on 10 high-rise residential buildings (with information on types of foundation, structure, exterior skins, SF, change orders, project delays, etc.) instead of inputting all 100 high-rise residential buildings we have completed to date, but with questionable data coming from the cost historical database.
5. Lastly, once the "smaller" database was ready, we pushed the actual entry of all future jobs in the database as part of the closeout process. In our department, we focused on "closing the project" from a schedule point of view, such as doing the final update with all actuals, leaving final thoughts on the delays/challenges and lastly, inputting everything in the scheduling historical database by the scheduler (one who knows most about the project schedule.)

Hope this helps. 

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.

Hi Dan, please read my paper DSAA-4200, which is presented at this year's ConEx... I believe I have provided an outline that answers your question(s)...

Does anybody in this community have experiencing standing up a Construction Data Analytics team?  Looking for some lessons learned specifically on the operational side rather than the IT side.