CM Pravda’s Post

In Building Information Modeling (BIM), the wealth of data at our fingertips has always fascinated me. Yet, it's disappointing that much of this rich information remains untapped. Since BIM's inception, the industry has shifted focus from 2D drawings to data, but there are no significant advancements in leveraging this data effectively. One major barrier is the expertise required to access and manipulate BIM data within complex authoring tools. Even after exporting to external databases, querying this data demands specialized knowledge or becomes cumbersome due to size, often deterring stakeholders from harnessing its full potential. I think this is where GenAI can step in. Imagine using natural language to query any aspect of a project, allowing stakeholders to extract valuable insights effortlessly. I've been experimenting with an idea where BIM models from Revit are exported into CosmosDB, with each project as a distinct collection. Using OpenAI's ada-002, I've created vector embeddings for these documents, enabling vector searches that surpass retrieving raw data. I am refining this approach, particularly addressing challenges such as optimizing aggregation queries where ChatGPT occasionally falls short. However, I'm close to a proof of concept and would love to hear from others who have explored similar strategies. #BIM #AI #ConstructionIndustry #GenAI #DataAnalytics #CosmosDB #ChatGPT #OpenAI

Wei Ni is spot on regarding leveraging BIM data. I've started doing the same kind of testing using Azure OpenAI Studio. It's blowing my mind at the ease of accessing the information. For me, the next step is ensuring that we get accurate information into the BIM datasets. AI will do all the heavy lifting once we get everything into the models. Thanks Wei!! #BIM #VDC #Construction #OpenAI

BIM and Data Science are always intertwined and deeply related. In addition, with the effects of artificial intelligence, the sector will move to much higher levels.

Another unfortunate point is that parts are designed in different design packages and then getting combined in packages like Navis works. Firstly, a lot of information will get lost during the transfer and second, there is no standard how information provided, is structured or named. This makes it almost impossible to use the otherwise great idea of BIM.

We can use all AI tools evolving the industry

I used ChatGPT to generate a 3D stair! I give width and height & it calculates all the steps for me. 🎩🪄 This is the episode 2 of my experiments to combine #openbim with #AI. The initial cube became a collection of funny shapes! If you missed the first episode, it's here, https://lnkd.in/dFRRMEi6 ╰┈➤ If you want to be following this experiment 🧐 Get notified if I make a tool that you can use. And even discuss what you think is an AI BIM editor Please sign up at this page https://lnkd.in/de2Hhx2G 🚩 So what I did that time ? Basically I played with the geometries to discover the key concepts of Three.js the library behind Open BIM component. You want to receive an email to be sure not to miss a step of these experiments. And get links to the experiments And I used #ChatGTP and GitHub Copilot to help me go fast. It does not replace programming, of course, as you need to understand the logic of the code. But it helps a lot, to start fast and then modify and refine the first version. So I could create : 1️⃣ A cube that even turn automatically on itself 2️⃣ A generator of a 3D cubes matrix with random colors ! 3️⃣ A stair generator : I give the height and width, it calculates the rest for me Great for three hours of works ! 💪 I struggled a bit with the camera that was hiding some of my objects depending on their scale. But finally I found a magic line of code to do a "zoom extents" and get all the scene visible. And finally I added a form input to be able to be able to write a prompt as "Stair please!" Or "stair just at the limit of the cube". That it could create me a future version. 🧠 This version of the experiments has no AI in it, but it's coming, stay tuned. The next step could be to interpret what the user want to "map" its demand to an existing generator (for example stair generator) and position it in 3D. And finally we may open even more to generate new shape on the fly, let's see! ╰┈➤ Live version here : https://lnkd.in/dxUwRsfG

Generative AI is affecting us all - learn more about how it is affecting the built environment and how you can apply it within your own organisation.

Artificial Intelligence is now a topic that comes up in nearly every conversation I have with people in our industry. There are so many questions surrounding it including what impact will it have to the job market, how will this affect project delivery, and what tools are available? One of the most common themes I hear is: how do I use it in my work? Since ChatGPT's introduction there has been an explosion in content and tools having this discussion, and Stjepan Mikulić has been at the forefront for our industry. He's created invaluable free resources helping us all get started on our AI journeys, and has just released his first course: Mastering AI in AEC, which will answer our burning questions about AI implementation. Stjepan has been a huge help to me in gaining perspective of AI in our industry, and am excited to start his course. If you're interested, sign up at the link below. I'm also interested in starting a study group for the course where we can discuss the content and learn from each other's experiences, so please reach out if you'd like to learn together! https://lnkd.in/e4exsQ4B #ai #bim #aec

AI security and the construction industry: learn how to avoid the risks and ensure a successful and safe integration.

𝐆𝐞𝐨𝐬𝐩𝐚𝐭𝐢𝐚𝐥 𝐡𝐚𝐬 𝐬𝐨 𝐦𝐚𝐧𝐲 𝐚𝐦𝐚𝐳𝐢𝐧𝐠 𝐬𝐭𝐨𝐫𝐢𝐞𝐬 𝐭𝐨 𝐭𝐞𝐥𝐥 .. 𝐛𝐮𝐭 𝐬𝐨𝐦𝐞𝐭𝐢𝐦𝐞𝐬 𝐭𝐡𝐚𝐭 𝐬𝐨𝐥𝐮𝐭𝐢𝐨𝐧 𝐬𝐭𝐨𝐫𝐲 𝐢𝐬 𝐛𝐞𝐬𝐭 𝐬𝐡𝐨𝐰𝐧 𝐚𝐧𝐝 𝐧𝐨𝐭 𝐭𝐨𝐥𝐝. The video below shows a cool new new innovative feature from Pix4D and their product PIX4Dcatch. The video/animation is really attention grabbing and informative (which is why I like video/animation so much). Now I have your attention (thanks to the video). Let's turn to words. Just as an exercise; I also wanted to describe this type of feature in simple value based terms. I asked ChatGPT to use Simon Sinek's approach (why, how, what). Note, this is me exploring the power of GenAI, I have altered the wording so it is generic and might be applied to this or any other competing product. Here is the result: Imagine the power of seeing your project's critical data overlaid onto the real world with perfect precision. 𝐖𝐡𝐲: Consistent, accurate mapping is crucial for effective project management, whether you're surveying land, managing construction, or mapping buried utilities. Precise ground control points (GCPs) are the backbone of reliable geospatial data. 𝐇𝐨𝐰: It is now possible to use augmented reality to transform how you visualize and capture GCPs. By overlaying digital 3D markers onto real-world scenes, you can be guided to the exact locations needed for high-accuracy captures, even across multiple sessions. This ensures your data is consistently precise, reducing errors and saving time. 𝐖𝐡𝐚𝐭: The AR Points features enhance your workflow by providing spatial guidance and ensuring that every GCP is captured accurately and efficiently. Whether you're scanning open trenches, visualizing subsurface utilities, or verifying construction projects. By integrating AR Points into your workflow, you can ensure that your projects are built on a foundation of accurate, reliable data, driving efficiency and reducing costs. Hmm, I like that. A powerful, attention grabbing video and compelling wording. What do you think? Credit to Pix4D for the video. If you are interested, here is a link to PIX4Dcatch: https://hubs.li/Q02yCSMZ0 𝐓𝐞𝐥𝐥𝐢𝐧𝐠 𝐲𝐨𝐮𝐫 𝐆𝐞𝐨𝐬𝐩𝐚𝐭𝐢𝐚𝐥 𝐒𝐭𝐨𝐫𝐲 𝐰𝐢𝐭𝐡 𝐆𝐞𝐧𝐀𝐈! Did this short post and animation catch your attention? This was all built with the help of GenAI. I believe this new tech can help geospatial companies better tell their story? Interested to learn more ... Let's chat. Download the eBook "𝐘𝐨𝐮𝐫 𝐂𝐨𝐦𝐩𝐥𝐞𝐭𝐞 𝐆𝐮𝐢𝐝𝐞 𝐭𝐨 𝐌𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 𝐆𝐞𝐧𝐀𝐈" to learn more about how to take your geospatial business to the next level: https://lnkd.in/ghJ457MU #GenAI #generativeai #marketing #sales #geospatial #spatial