Hello, everyone. It's a pleasure to present to you today. My name is Scott Johnson. I am a data systems manager in POET, which is in Sioux Falls, South Dakota.

I'm Brandon Breitling, senior statistician at POET in Sioux Falls as well.

POET is the world's largest producer of bioethanol and our headquarters is in Sioux Falls, South Dakota, I mentioned. We have 34 bioprocessing sites across eight states in the Midwest. We have a unique integrated business model where POET manages site selection, design and construction, grain sourcing, product marketing, plant management. We even get into public policy within related AG industries.

As such, it can be challenging to deliver data and insights to the right people at the right time. Of course, this is not a unique opportunity at POET. Many companies have similar data analysis challenges. Like other companies, POET uses many different tools to deliver reports, KPIs, dashboards. We have a great need, in particular for real-time data analysis.

The core of our business is done at our 34 bioprocessing sites that run 24/7. JMP has been one of the key platforms that help us deliver those insights to users across our company. Rather than a real specific JMP use case, we want to show some examples of how we use JMP at a corporate or fleet support level, where we do all the hard analysis and chart building at that fleet level, and how then we share those findings with others.

We will demonstrate how adding JMP Live recently to our tool set has opened up even greater opportunities to share findings across the network. I'm going to pass it over to Brandon to show us some of those examples.

Wonderful. Thanks, Scott. JMP has been a tool that POET has used for going back more than a decade to analyze trends, design experiments, and query data. I help to manage an add-in for JMP with a screenshot in the left here. The add-in points to a large number of tables and views stored in POET's data warehouse, where on a regular cadence, data is queried for multiple transactional systems and is integrated and stored or access from the add-in.

These data sets are then available for analysis for engineers, scientists, and business people. I focus a lot of my time helping to set up the data sources with the help of primary data owners to extract, transfer, and load the data from more raw or transactional sources.

I then model the data into useful sets for insights for the end users. Generally, it's more prudent for the organization if I do these steps in the warehouse, but some of that data wrangling and modeling can be done with JSL as well. Looking back to the add-in screenshot, the upper left here, if you look to the leftmost pane, headers like fermentation, if you go down a little bit, distillation, evaps, these are ethanol production unit operations, and much of the data in the add-in is operations-focused. Most of that operations transactional data is housed in OSI PI, which Scott on the call here and folks on his team are the primary data handlers for.

JMP has a way to directly connect to OSI PI, seen in the screenshot in the upper right back in the presentation. When pulling larger data sets, especially for multiple sources or with calculated columns for this data, I work with Scott and his team to get that raw data into the warehouse where I then model it for end users. We also have a lot of financial and other transactional systems that we can integrate with operations for making insight-ready data sets for analysis.

Shown in the lower right screenshot, users can do queries and joins directly if they know the tables and the keys. They can either pull add-in tables and go to join under the table's menu with JMP data sets open on the desktop, or they can query, again, similar to that screenshot, directly to the database itself using the database's platform under the file menu, and select Databases and then Direct Queering tables that haven't been linked in the add-in yet.

As an extension to the many JMP licenses we have, I currently have a JMP Pro license. I use it for more specific modeling requests that may leverage the enhanced capabilities of JMP Pro. For a lot of operations data, the predictors are correlated. If I need to do a regression model shown to the left, generalized regression, LASSO is appropriate due to the correlated predictors, and it is a JMP Pro feature.

Also, our fermentations are done batch by batch, and the kinetics themselves are curves, which are appropriate responses for Functional Data Explorer platform. With a snapshot to the right, we have an example. Batch level predictors for differences in the kinetics can be explored with…

Scott will go into more detail on specific JMP Live use case. With JMP Live, we have a lot of capabilities similar to other business intelligence tools that allow for interactivity with charts and analysis, but a few unique capabilities make it superior for certain tasks.

It has a robust quality control charting, which an example is shown in the chart with the green in the middle and the yellow surrounding that and the red on the outermost, the upper left chart. If you look in that chart, too, there's a capability analysis to the right with the histograms and CPKs and some other statistics. Those are all easy to add. You can also add warnings, too, and then they can send out emails automatically when the criteria are met.

Also in the lower left here, it is very easy to plot data on a background map to investigate for geographic-related insights. More traditional bar graph-type insights can be generated too, with some analytic querying abilities using the data filters which are blocked out in the column switchers, which are seen on the right. These JMP Live analysis can be easily accessed by navigating to the portal and clicking on any one of the snippets shown above.

With that, I will transfer back to Scott for some more discussion and more specific JMP Live use case.

Thank you, Brandon. With JMP Live added to our tool set, it was about a year and a half ago or so, I would say we now have three levels of JMP users within POET, and that's somewhat reflected in the actual JMP software versions or tools. That is the JMP client in our engineers' research, subject-matter experts. They utilize the data add-in that Brandon highlighted earlier for that easy access to the data.

Again, the smaller subset using JMP Pro to deeper dive into the data, and it's a very small subset of our users. Now we're able to expose through JMP Live, which is new for POET. This allows us to take advantage of some of those findings that our engineers and subject-matter experts have found, and then easily publish those screens to a bigger audience.

We'll focus a little bit on JMP Live and talk about a case around fermentation, which actually was a specific use case that drove us to adding JMP Live to our tool set. A key part of the bioethanol process is fermentation, and we've used JMP and JMP Pro in the last for some one-off analysis. Those findings were only easily shared with other users who had JMP installed and were familiar with how to use software.

This somewhat limited accessibility was familiar with some of the other tools we had. It seemed that either the tool did not have robust flexibility for us to easily drill down and troubleshoot issues, or the tool had that power, but it was not easily accessible to a larger audience. We need that larger audience to help make quick decisions to mitigate any problems within fermentation.

Here's an example of a file we built. This is using SQL Server reporting services to watch fermentation at any of our plants. It's a curve of ethanol conversion over time and batches overlaid on top of each other. This view did allow us to search by plant and by time range, but it lacked a lot of the power of drill-down capability that we're looking for. Couldn't easily go very deep into the data without subject matter expertise or someone mining the data in somewhat manual ways.

We were able to build this out in JMP client tool and similar views with that, and it was very useful in slicing, dicing, drilling down the data. That analysis, again, was only limited to our JMP users. It was not easily shared within a large audience. Those with access to JMP were the only ones that could take advantage of them.

Enter JMP Live. We were able to essentially export that functionality that we had in the JMP Client tool out, and publish it to JMP Live, where it is now accessible to our fleet support. A couple of the screenshots here highlight some of the benefits of that. In the upper left-hand corner, we can see where one of the fermenters, one of the batches went off the rails, so to speak. It's that green line that has a lower ethanol conversion rate over time.

Using JMP Live, if we dig deeper, if people want to drill into that and find out what was going on, the users can highlight any outliers and then see what went into that. In this case, we can see the batch number, the ferm tank number, the vessel that was in. Interestingly for us, it can tell us what recipe strategies we were using at the time, which yeast we were using, which enzyme we were using.

In the upper right-hand corner, speaking of those yeast and enzymes, we can highlight a particular enzyme and see if there were any leading indicators in previous batches. I highlighted some sections that might indicate that this strategy was going to be lower performing.

In the lower right, highlights some of that same data, but different views of that data set. It highlights that same plant where we had an underperforming batch, but then it highlights it against the rest of the fleet, other plants, and we can see what strategies they're using. That's a very powerful and common tool that we look for in our software platforms at POET, the ability to compare against our fleet and then use those strategies of the high performers, try to raise the bar for the rest of the group.

Just an indicator adds to the success of this use case, these fermentation views, this particular one, has over 15,000 views across our group since we implemented it about a year and a half ago. It's accessed by hundreds of users across their fleet.

Hopefully, this gives you some insight and illustrates examples of how JMP is playing a pivotal role in our data-driven decision-making. It helps us with process optimization, quality assurance initiatives, and so forth within our organization.

As we continue to enhance the access to some of those data sets that Brandon mentioned earlier at an enterprise level, and now with the addition of JMP Live to our tool set, that shows us that the JMP platform is becoming an indispensable ally in our pursuit of operational excellence and continuous improvement.

Thank you for watching, and have a good day.

Thank you.