Top 5 ways Locus Water Quality Monitoring Software helps utility companies resolve challenges with centralized, reliable, & actionable data.
Find out why AI EHS compliance software only works properly in a technical environment architected for artificial intelligence models.
Locus Technologies continues to explore new frontiers such as a “natural language” AI chatbot for data stored in Locus software.
There are two promising technologies that are about to change how we aggregate and manage EHS+S data: artificial intelligence (AI) and blockchain. When it comes to technology, history has consistently shown that the cost will always decrease, and its impact will increase over time. We still lack access to enough global information to allow AI to make a significant dent in global greenhouse gas (GHG) emissions by merely providing better tools for emissions management. For example, the vast majority of energy consumption is wasted on water treatment and movement. AI can help optimize both. Along the way, water quality management becomes an add-on app.
AI is a collective term for technologies that can sense their environment, think, learn, and act in response to what they’re detecting and their objectives. Possible applications include (1) Automation of routine tasks like sampling and analyses of water samples, (2) Segregation of waste disposal streams based on the waste containers content, (3) Augmentation of human decision-making, and (4) Automation of water treatment systems. AI systems can greatly aid the process of discovery – processing and analyzing vast amounts of data for the purposes of spotting and acting on patterns, skills that are difficult for humans to match. AI can be harnessed in a wide range of EHS compliance activities and situations to contribute to managing environmental impacts and climate change. Some examples of applications include permit interpretation and response to regulatory agencies, precision sampling, predicting natural attenuation of chemicals in water or air, managing sustainable supply chains, automating environmental monitoring and enforcement, and enhanced sampling and analysis based on real-time weather forecasts. Applying AI in water resource prediction, management, and monitoring can help to ameliorate the global water crisis by reducing or eliminating waste, as well as lowering costs and lessening environmental impacts. A similar analogy holds for air emissions management.
The onset of blockchain technology will have an even bigger impact. It will first liberate data and, second, it will decentralize monitoring while simultaneously centralizing emissions management. It may sound contradictory, but we need to decentralize in order to centralize management and aggregate relevant data across corporations and governmental organizations without jeopardizing anyone’s privacy. That is the power of blockchain technology. Blockchain technology will eliminate the need for costly synchronization among stakeholders: corporations, regulators, consultants, labs, and the public. What we need is secure and easy access to any data with infinite scalability. It is inevitable that blockchain technology will become more accessible with reduced infrastructure over the next few decades. My use of reduced architecture here refers to a replacement of massive centralized databases controlled by one of the big four internet companies using the hub-and-spoke model concept with a device-to-device communication with no intermediaries.
This post was originally published in Environmental Business Journal in June of 2020.
Locus Founder and CEO, Neno Duplan recently sat down with Grant Ferrier of the Environmental Business International to discuss a myriad of topics relating to technology in the environmental industry such as Artificial Intelligence, Blockchain, Multi-tenancy, IoT, and much more.
In August 2014, we wrote on the potential use of wearables for EHS professionals. Less than a year later, the Apple Watch was introduced, revolutionizing the market. Now, wearables in the EHS space aren’t a hypothetical. Roughly a fifth to a quarter of Americans wear a smartwatch daily. Wearables are undoubtedly one of the biggest trends in EHS, with a seemingly endless number of uses to promote a more efficient and safer workplace.
Despite recent growth, wearables are still in their infancy when it comes to EHS. Verdantix anticipates that companies will spend 800% more on connected worker devices in twenty years, an explosion in utilization. This year alone, over 20% of surveyed companies are reporting an increase in budget for wearables for EHS purposes. While demand from organizations is growing, most EHS software is yet to adapt to market needs, with few offering wearable support.
Locus is prepared to meet the needs of the market, by integrating wearable support with our mobile application. Here are a few ways to best utilize your smartwatch with Locus Mobile:
Custom and priority notifications can be tailored to fit the needs of professionals in your organization, increasing engagement and response time.
Calendar alerts directly to your wearable, so that no samples are missed by field technicians.
Get alerted when you’re entering a safety zone that requires specific PPE.
Track worker vital signs for faster response time in the event of an emergency.
If your organization is looking to implement wearable tech, Locus product specialists are ready to discuss your needs and how we can help.
The recent year of lockdowns pushed many daily activities into the virtual world. Work, school, commerce, the arts, and even medicine have moved online and into the cloud. As a result, considerably more resources and information are now available from an internet browser or from an application on a handheld device. To navigate through all this content and make sense of it, you need the ability to quickly search and get results that are most relevant to your needs.
You can think of the web as a big database in the cloud. Traditionally, database searches were done using a precise syntax with a standard set of keywords and rules, and it can be hard for non-specialists to perform such searches without learning programming languages. Instead, you want to search in as natural a matter as possible. For example, if you want to find pizza shops with 15 miles of your house that offer delivery, you don’t want to write some fancy statement like “return pizza_shop_name where (distance to pizza shop from my house < 15 miles) and (offers_delivery is true). You just want to type “what pizza shops within 15 miles of my house offer delivery?” How can this be done?
Enter the search engine. While online search engines appeared as early as 1990, it wasn’t until Yahoo! Search appeared in 1995 that their usage became widespread. Other engines such as Magellan, Lycos, Infoseek, Ask Jeeves, and Excite soon followed, though not all of them survived. In 1998, Google hit the internet, and it is now the most dominant engine in use. Other popular engines today are Bing, Baidu, and DuckDuckGo.
Current search engines compare your search terms to proprietary indexes of web page and their content. Algorithms are used to determine the most relevant parts of the search terms and how the results are ranked on the page. Your search success depends on what search terms you enter (and what terms you don’t enter). For example, it is better to search on ‘pizza nearby delivery’ than ‘what pizza shops that deliver are near my house’, as the first search uses less terms and thus more effectively narrows the results.
Search engines also support the use of symbols (such as hyphens, colons, quote marks) and commands (such as ‘related’, ‘site’, or ‘link’) that support advanced searches for finding exact word matches, excluding certain results, or limiting your search to certain sites. To expand on the pizza example, support you wanted to search for nearby pizza shops, but you don’t want to include Nogud Pizza Joints because they always put pineapple on your pizza. You would need to enter ‘pizza nearby delivery -nogud’. In some ways, with the need to know special syntax, searching is back where it was in the old database days!
Search engines are also a key part of ‘digital personal assistants’, or programs that not only perform searches but also perform simple tasks. An assistant on your phone might call the closest pizza shop so you can place an order, or perhaps even login to your loyalty app and place the order for you. There is a dizzying array of such assistants used within various devices and applications, and they all seem to have soothing names such as Siri, Alexa, Erica, and Bixby. Many of these assistants support voice activation, which just reinforces the need for natural searches. You don’t want to have to say “pizza nearby delivery minus nogud”! You just want to say “call the nearest pizza shop that does delivery, but don’t call Nogud Pizza”.
Search engine and digital personal assistant developers are working towards supporting such “natural” requests by implementing “natural language processing”. Using natural language processing, you can use full sentences with common words instead of having to remember keywords or symbols. It’s like having a conversation as opposed to doing programming. Natural language is more intuitive and can help users with poor search strategies to have more successful searches.
Furthermore, some engines and assistants have artificial intelligence (AI) built in to help guide the user if the search is not clear or if the results need further refinement. What if the closest pizza shop that does delivery is closed? Or what if a slightly farther pizza place is running a two-for-one special on your favorite pizza? The built-in AI could suggest choices to you based on your search parameters combined with your past pizza purchasing history, which would be available based on your phone call or credit charge history.
The Locus team recently expanded the functionality of the EIM (Environmental Information Management) search bar to support different types of data searches. If a search term fits several search types, all are returned for the user to review.
EIM lets the user perform successful searches through various methods. In all searches, the user does not need to specify if the search term is a menu item, help page, or data entity such as parameter or location. Rather, the search bar determines the most relevant results based on the data currently in EIM. Furthermore, the search bar remembers what users searched for before, and then ranks the results based on that history. If a user always goes to a page of groundwater levels when searching for location ‘MW-1’, then that page will be returned first in the list of results. Also, the EIM search bar supports common synonyms. For example, searches for ‘plot’, ‘chart’, and ‘graph’ all return results for EIM’s charting package.
By implementing the assistance methods described above, Locus is working to make searching as easy as possible. As part of that effort, Locus is working to add natural language processing into EIM searches. The goal is to let users conduct searches such as ‘what wells at my site have benzene exceedances’ or perform tasks such as ‘make a chart of benzene results’ without having to know special commands or query languages.’
How would this be done? Let’s set aside for now the issues of speech recognition – sadly, you won’t be talking to EIM soon! Assume your search query is ‘what is the maximum lead result for well 1A?’
A similar process could be done for tasks such as ‘make a chart of xylene results’. In this case, however, there is too much ambiguity to proceed, so EIM would need to return queries for additional clarifications to help guide the user to the desired result. Should the chart show all dates, or just a certain date range? How are non-detects handled? Which locations should be shown on the chart? What if the database stores separate results for o-Xylene, m,p-Xylene, plus Xylene (total)? Once all questions were answered, EIM could generate a chart and return it to the user.
Natural language is the key to helping users construct effective searches for data, whether in EIM, on a phone, or in the internet. Locus continues to improve EIM by bringing natural language processing to the EIM search engine.
About the Author—Dr. Todd Pierce, Locus Technologies
Dr. Pierce manages a team of programmers tasked with development and implementation of Locus’ EIM application, which lets users manage their environmental data in the cloud using Software-as-a-Service technology. Dr. Pierce is also directly responsible for research and development of Locus’ GIS (geographic information systems) and visualization tools for mapping analytical and subsurface data. Dr. Pierce earned his GIS Professional (GISP) certification in 2010.
Regardless of the size of your organization or the industry you’re in, chances are that right now artificial intelligence can benefit your EHS&S initiatives in one way or another. And whether you are ready for it or not, the age of artificial intelligence is coming. Forward-thinking and adaptive businesses are already using artificial intelligence in EHS&S as a competitive advantage in the marketplace to great success.
With modern EHS&S software, immense amounts of computing power, and seemingly endless cloud storage, you now have the tools to achieve fully-realized AI for your EHS&S program. And while you may not be ready to take the plunge into AI just yet, there are some steps you can take to implement artificial intelligence into your EHS&S program in the future.
Perhaps the best aspect of preparing for AI implementation is that all of the steps you take to properly bring about an AI system will benefit your program even before the deployment phase. Accurate sources, validated data, and one system of record are all important factors for any EHS&S team.
Used alongside big data, AI can quickly draw inferences and conclusions about many aspects of life more efficiently than with human analysis, but only if your sources pull accurate data. Accurate sources data will help your organization regardless of your current AI usage level. That’s why the first step to implementing artificial intelligence is auditing your data sources.
Sources pulling accurate data can be achieved with some common best practices. First, separate your data repository from the process that analyzes the data. This allows you to repeat the same analysis on different sets of data without the fear of not being able to replicate the process of analysis. AI requires taking a step away from an Excel-based or in-house software, and moving to a modern EHS&S software, like Locus Platform that will audit your data as it is entered. This means that anything from SCADA to historical outputs, samples, and calculations can be entered and vetted. Further, consider checking your data against other sources and doing exploratory analysis to greater legitimize your data.
AI requires data, and a lot of it—aggregated from multiple sources. But no amount of predictive analysis or machine learning is going to be worth anything without proper data validation processes.
Collected data must be relevant to the problem you are trying to solve. Therefore, you need validated data, which is a truly difficult ask with Excel, in-house platforms, and other EHS&S software. Appropriate inputs, appropriate ranges, data consistency, range checks (to name a few)—are all aspects of data that is validated in a modern EHS&S software like Locus Platform. Without these checks inherent to a platform, you cannot be sure that your data, or your analyses are producing useful or accurate results.
Possibly the best reason to get started with AI is the waterfall effect. As your data uncovers hidden insights and starts to learn on its own, the more accurate your new data will be and the better your predictions will become.
A unified system of record and a central repository for all data means that you see an immediate increase in data quality. Starting with AI means the end of disconnected EHS&S systems. No more transferring data from one platform to another or from pen and paper, just fully-digitized and mobile-enabled data in one platform backed up in the cloud. You also gain the added benefit of being able to access your data in real-time, incorporate compliance/reporting on the fly, and save time and resources using a scalable solution instead of a web of spreadsheets and ad-hoc databases.
Whether you are ready for AI or not, investing in these otherwise useful steps are necessary for any program looking to harness the power of artificial intelligence. When you are ready to take that next step, you will be well on the path to AI implementation, with a solid data infrastructure in place for your efforts.
To learn more about artificial intelligence, view this NAEM-hosted webinar led by Locus experts, or read our study on predicting water quality using machine learning.
At Locus Technologies, we’re always looking for innovative ways to help water users better utilize their data. One way we can do that is with powerful technologies such as machine learning. Machine learning is a powerful tool which can be very useful when analyzing environmental data, including water quality, and can form a backbone for competent AI systems which help manage and monitor water. When done correctly, it can even predict the quality of a water system going forward in time. Such a versatile method is a huge asset when analyzing data on the quality of water.
To explore machine learning in water a little bit, we are going to use some groundwater data collected from Locus EIM, which can be loaded into Locus Platform with our API. Using this data, which includes various measurements on water quality, such as turbidity, we will build a model to estimate the pH of the water source from various other parameters, to an error of about 1 pH point. For the purpose of this post, we will be building the model in Python, utilizing a Jupyter Notebook environment.
When building a machine learning model, the first thing you need to do is get to know your data a bit. In this case, our EIM water data has 16,114 separate measurements. Plus, each of these measurements has a lot of info, including the Site ID, Location ID, the Field Parameter measured, the Measurement Date and Time, the Field Measurement itself, the Measurement Units, Field Sample ID and Comments, and the Latitude and Longitude. So, we need to do some janitorial work on our data. We can get rid of some columns we don’t need and separate the field measurements based on which specific parameter they measure and the time they were taken. Now, we have a datasheet with the columns Location ID, Year, Measurement Date, Measurement Time, Casing Volume, Dissolved Oxygen, Flow, Oxidation-Reduction Potential, pH, Specific Conductance, Temperature, and Turbidity, where the last eight are the parameters which had been measured. A small section of it is below.
Alright, now our data is better organized, and we can move over to Jupyter Notebook. But we still need to do a bit more maintenance. By looking at the specifics of our data set, we can see one major problem immediately. As shown in the picture below, the Casing Volume parameter has only 6 values. Since so much is missing, this parameter is useless for prediction, and we’ll remove it from the set.
We can check the set and see that some of our measurements have missing data. In fact, 261 of them have no data for pH. To train a model, we need data which has a result for our target, so these rows must be thrown out. Then, our dataset will have a value for pH in every row, but might still have missing values in the other columns. We can deal with these missing values in a number of ways, and it might be worth it to drop columns which are missing too much, like we did with Casing Volume. Luckily, none of our other parameters are, so for this example I filled in empty spaces in the other columns with the average of the other measurements. However, if you do this, it is necessary that you eliminate any major outliers which might skew this average.
Once your data is usable, then it is time to start building a model! You can start off by creating some helpful graphs, such as a correlation matrix, which can show the relationships between parameters.
For this example, we will build our model with the library Keras. Once the features and targets have been chosen, we can construct a model with code such as this:
This code will create a sequential deep learning model with 4 layers. The first three all have 64 nodes, and of them, the initial two use a rectified linear unit activation function, while the third uses a sigmoid activation function. The fourth layer has a single node and serves as the output.
Our model must be trained on the data, which is usually split into training and test sets. In this case, we will put 80% of the data into the training set and 20% into the test set. From the training set, 20% will be used as a validation subset. Then, our model examines the datapoints and the corresponding pH values and develops a solution with a fit. With Keras, you can save a history of the reduction in error throughout the fit for plotting, which can be useful when analyzing results. We can see that for our model, the training error gradually decreases as it learns a relationship between the parameters.
The end result is a trained model which has been tested on the test set and resulted in a certain error. When we ran the code, the test set error value was 1.11. As we are predicting pH, a full point of error could be fairly large, but the precision required of any model will depend on the situation. This error could be improved through modifying the model itself, for example by adjusting the learning rate or restructuring layers.
You can also graph the true target values with the model’s predictions, which can help when analyzing where the model can be improved. In our case, pH values in the middle of the range seem fairly accurate, but towards the higher values they become more unreliable.
So what do we do now that we have this model? In a sense, what is the point of machine learning? Well, one of the major strengths of this technology is the predictive capabilities it has. Say that we later acquire some data on a water source without information on the pH value. As long as the rest of the data is intact, we can predict what that value should be. Machine learning can also be incorporated into examination of things such as time series, to forecast a trend of predictions. Overall, machine learning is a very important part of data analytics and the development of powerful AI systems, and its importance will only increase in the future.
As the technology around machine learning and artificial intelligence evolves, Locus will be working to integrate these tools into our EHS software. More accurate predictions will lead to more insightful data, empowering our customers to make better business decisions.
With data and information streaming from devices like fire hydrants, there is little benefit from raw data, unless a company owning the data has a way to integrate the data into its record system and pair it with regulatory databases and GIS. That is where the advancement in SaaS tools and data sources mashups has helped set the stage for AI as a growing need.
Humans are not very good at analyzing large datasets. This is particularly true with data at the planetary level that are now growing exponentially to understand causes and fight climate change. Faced with a proliferation of new regulations and pressure to make their companies “sustainable” EHS departments keep adding more and more compliance officers, managers, and outside consultants, instead of investing in technology that can help them. Soon, they will be turning to AI technology to stay on top of the ever-changing regulatory landscape.
AI, in addition to being faster and more accurate, should make compliance easier. Companies spend too much time and effort on the comprehensive quarterly or annual reporting—only to have to duplicate the work for the next reporting period. The integrated approach, aided by AI, will automate these repetitive tasks and make it easier than just having separate analyses performed on every silo of information before having a conversation with regulators.
In summary, whether it is being used to help with GHG emissions monitoring and reporting, water quality management, waste management, incident management, or other general compliance functions, AI can improve efficiency, weed out false-positive results, cut costs and make better use of managers’ time and company resources.
Another advantage of AI, assuming it is deployed properly, concerns its inherent neutrality on data evaluation and decision making. Time and time again we read in the papers about psychological studies and surveys that show people on opposite sides of a question or topic cannot even agree on the “facts.” It should not be surprising then to find that EHS managers and engineers are often limited by their biases. As noted in the recent best-seller book by Nobel Memorial Prize in Economics laureate Daniel Kahneman, “Thinking, Fast and Slow,” when making decisions, they frequently see what they want, ignore probabilities, and minimize risks that uproot their hopes. Even worse, they are often confident even when they are wrong. Algorithms with AI built-in are more likely to detect our errors than we are. AI-driven intelligent databases are now becoming powerful enough to help us reduce human biases from our decision-making. For that reason, large datasets, applied analytics, and advanced charting and data visualization tools, will soon be driving daily EHS decisions.
In the past, companies almost exclusively relied upon on-premise software (or single-tenant cloud software, which is not much different from on-premise). Barriers were strewn everywhere. Legacy systems did not talk to one another, as few of the systems interfaced with one another. Getting data into third-party apps usually required the information to be first exported in a prescribed format, then imported to a third-party app for further processing and analysis. Sometimes data was duplicated across multiple systems and apps to avoid the headache of moving data from one to another. As the world moves to the multi-tenant SaaS cloud, all this is now changing. Customers are now being given the opportunity to analyze not just their company’s data, but data from other companies and different but potentially related and coupled categories via mashups. As customers are doing so, interesting patterns are beginning to emerge.
The emergence of artificial intelligence is a game-changer for enterprise EHS and content management because it can deliver business insights at scale and make EHS compliance more productive. There are numerous advantages when you combine the leading multi-tenant EHS software with AI:
We are seeing in the most recent NAEM white paper, Why Companies Replace Their EHS&S Software Systems, that people want the ability to integrate with other systems as a top priority. Once the ability to share/consolidate data is available, AI is not far behind in the next generation of EHS/Water Quality software.
This concludes the four-part blog series on Big Data, IoT, AI, and multi-tenancy. We look forward to feedback on our ideas and are interested in hearing where others see the future of AI in EHS software – contact us for more discussion or ideas! Read the full Series: Part One, Part Two, Part Three.
Neno Duplan is founder and CEO of Locus Technologies, a Silicon Valley-based environmental software company founded in 1997. Locus evolved from his work as a research associate at Carnegie Mellon in the 1980s, where he developed the first prototype system for environmental information management. This early work led to the development of numerous databases at some of the nation’s largest environmental sites, and ultimately, to the formation of Locus in 1997.
Mr. Duplan recently sat down with Environmental Business Journal to discuss a myriad of topics relating to technology in the environmental industry such as Artificial Intelligence, Blockchain, Multi-tenancy, IoT, and much more.
Click here to learn more and purchase the full EBJ Vol XXXIII No 5&6: Environmental Industry Outlook 2020-2021